{ "format_version": 1, "title": "Frozen offline installed-options evidence", "publication_status": "sanitized_evidence_export", "snapshot_id": "20261010T224600Z-all-installed-options-interrupted-evidence", "snapshot_created_at_original_offset": "2026-10-10T16:46:01.864254-06:00", "scope": "Frozen original-v3 plus independently sealed IQ2 tail4 and Q4 tail9, and diagnostic interrupted IQ2 records. This is not a new inference run.", "study_ids": { "original-v3": "20261009T192800Z-all-installed-options-checkpoint-v3", "iq2-interrupted-missing14": "20261010T040600Z-iq2-missing-cells-supplement", "iq2-interrupted-observation": "20261010T224100Z-iq2-interrupted-observation", "iq2-fresh-tail4": "20261010T224200Z-iq2-tail4-recovery", "q4-tail9": "20261010T221520Z-q4-tail9-supplement" }, "gates": { "provenance_verified": true, "recorded_termination_verified": false, "all_selected_lifecycles_qualified": false, "live_machine_state_checked": false, "combined_attempted_or_intent_coverage_complete": true, "combined_response_record_coverage_complete": false, "original_v3_benchmark_execution_complete": false, "all_options_benchmark_execution_complete": false, "iq2_tail4_benchmark_execution_complete": true, "q4_tail9_benchmark_execution_complete": true, "product_acceptance": "pending", "qa_grades_imported": false }, "counts": { "planned_canonical_targets": 179, "planned_generation_targets": 174, "planned_embedding_targets": 5, "attempted_or_intent_canonical_targets": 179, "response_recorded_canonical_targets": 178, "unattempted_targets": 0, "unfinalized_pending_targets": 1, "original_response_records": 156, "interrupted_iq2_response_records": 9, "interrupted_iq2_persisted_intents": 10, "fresh_tail_response_records": 13, "diagnostic_only_response_records": 20, "formal_complete_response_deliveries": 110, "observed_complete_response_deliveries_including_diagnostics": 128 }, "cell_count_cross_check": { "cells": 179, "attempted_or_intent": 179, "response_recorded": 178, "pending": 1, "diagnostic_response_records": 20, "formal_complete": 110, "observed_complete": 128 }, "recorded_call_ledger": { "original_selected_generation_calls": 151, "original_selected_embedding_calls": 5, "original_all_attempts_recorded_generation_calls": 153, "original_all_attempts_recorded_embedding_calls": 5, "excluded_noncanonical_recorded_generation_calls": 2, "excluded_noncanonical_recorded_embedding_calls": 0, "interrupted_iq2_persisted_generation_intents": 10, "interrupted_iq2_persisted_response_records": 9, "interrupted_iq2_pending_intent_dispatch_verified": false, "iq2_tail4_recorded_generation_calls": 4, "q4_tail9_recorded_generation_calls": 9, "all_studies_persisted_attempt_or_intent_counters": 181, "original_declared_embedding_load_unload_operations": 2, "limits": "Persisted intent/call counters only; pending dispatch unverified; not usage, cost or successful responses." }, "excluded_noncanonical_attempts": { "candidate_id": "strata-ud-q4_k_xl", "attempt": 1, "state": "started", "started_at_utc": "2026-10-09T16:41:05.822153-06:00", "generation_calls_attempted": 2, "embedding_calls_attempted": 0 }, "frozen_environment": { "gpu": "NVIDIA GeForce RTX 5080, 617.14, 16303 MiB", "inventory_ollama_version": "0.40.2", "measured_owned_ollama_server_version": "0.40.2", "ollama_version_scope": "Inventory and all six selected owned candidate server logs independently record 0.40.2. Protocol references to official 0.40.1 source explain native name behavior; they are not the measured runtime version.", "strata_engine_version": "0.1.41", "physical_memory_bytes": 66202144768, "system_commit_limit_gib": 65.53055191040039, "configured_context_tokens": 32768, "configured_parallelism": 1, "thresholds": { "minimum_sampled_available_physical_gib": 4, "minimum_sampled_available_system_commit_gib": 2, "minimum_samples_each_scope": 1, "maximum_sampling_errors_each_scope": 0 }, "generation_timeout_seconds": 1800, "load_timeout_seconds": 900, "fixed_order": [ "ollama-gptoss-ggml", "ollama-gptoss-llamacpp", "ollama-qwen35-ggml", "ollama-qwen35-llamacpp", "ollama-qwen3vl", "ollama-embedding", "strata-iq2_xs", "strata-iq3_xxs", "strata-iq3_s", "strata-ud-q4_k_xl" ], "asset_identity_limits": "Original defaults/source-store size+mtime and manifest hashes; exact private initial manifest SHA. Private runtime cache/migrations permitted. No full rehash of large original weights in this inventory." }, "candidates": [ { "candidate_id": "ollama-gptoss-ggml", "backend": "ollama", "model_id": "gpt-oss:20b", "frozen_request_model_name": "ggml:21fd97703fbeb8b4a74b2ef1648aca44c8daeb32dcb019f88ec5ab4c97397892", "kind": "text", "manifest_digest": "21fd97703fbeb8b4a74b2ef1648aca44c8daeb32dcb019f88ec5ab4c97397892", "manifest_label": "ggml", "installed_size_bytes": 13793441244, "parameter_size": "20.9B", "quantization": "MXFP4", "configured_context_tokens": 32768, "kv_policy": "native installed defaults, not claimed equivalent to Strata int8", "configured_parallelism": 1, "has_vision_fixture": false }, { "candidate_id": "ollama-gptoss-llamacpp", "backend": "ollama", "model_id": "gpt-oss:20b", "frozen_request_model_name": "llamacpp:9ba9cc2b4e02463b933096090530aef679d534dd91ecc9afadfb2dbfdc59f3c5", "kind": "text", "manifest_digest": "9ba9cc2b4e02463b933096090530aef679d534dd91ecc9afadfb2dbfdc59f3c5", "manifest_label": "llamacpp", "installed_size_bytes": 13793441546, "parameter_size": "20.9B", "quantization": "MXFP4", "configured_context_tokens": 32768, "kv_policy": "native installed defaults, not claimed equivalent to Strata int8", "configured_parallelism": 1, "has_vision_fixture": false }, { "candidate_id": "ollama-qwen35-ggml", "backend": "ollama", "model_id": "qwen3.5:9b", "frozen_request_model_name": "ggml:2e16a80fe3d77d431d295466415848403c8c266667ee5d7d8991ee445b4e1311", "kind": "text", "manifest_digest": "2e16a80fe3d77d431d295466415848403c8c266667ee5d7d8991ee445b4e1311", "manifest_label": "ggml", "installed_size_bytes": 6594474711, "parameter_size": "9.7B", "quantization": "Q4_K_M", "configured_context_tokens": 32768, "kv_policy": "native installed defaults, not claimed equivalent to Strata int8", "configured_parallelism": 1, "has_vision_fixture": true }, { "candidate_id": "ollama-qwen35-llamacpp", "backend": "ollama", "model_id": "qwen3.5:9b", "frozen_request_model_name": "llamacpp:c97eb11d70b1acdc88af01eef566c1fe4f7fbe93eb1afc06871132f293ff425a", "kind": "text", "manifest_digest": "c97eb11d70b1acdc88af01eef566c1fe4f7fbe93eb1afc06871132f293ff425a", "manifest_label": "llamacpp", "installed_size_bytes": 6603314868, "parameter_size": "9.7B", "quantization": "Q4_K_M", "configured_context_tokens": 32768, "kv_policy": "native installed defaults, not claimed equivalent to Strata int8", "configured_parallelism": 1, "has_vision_fixture": true }, { "candidate_id": "ollama-qwen3vl", "backend": "ollama", "model_id": "qwen3-vl:8b-instruct", "frozen_request_model_name": "ggml:0533d74300e4f9bc367d675d4e64ffd073d50ff16a2b4096cc2e8a1cf8c96319", "kind": "text", "manifest_digest": "0533d74300e4f9bc367d675d4e64ffd073d50ff16a2b4096cc2e8a1cf8c96319", "manifest_label": "ggml", "installed_size_bytes": 6140415975, "parameter_size": "8.8B", "quantization": "Q4_K_M", "configured_context_tokens": 32768, "kv_policy": "native installed defaults, not claimed equivalent to Strata int8", "configured_parallelism": 1, "has_vision_fixture": true }, { "candidate_id": "ollama-embedding", "backend": "ollama", "model_id": "qwen3-embedding:0.6b", "frozen_request_model_name": "ggml:ac6da0dfba84a81fdbfbaf330198c33cd77c4cdfc53e8bc50eb581914a15621d", "kind": "embedding", "manifest_digest": "ac6da0dfba84a81fdbfbaf330198c33cd77c4cdfc53e8bc50eb581914a15621d", "manifest_label": "ggml", "installed_size_bytes": 639150858, "parameter_size": "595.78M", "quantization": "Q8_0", "configured_context_tokens": 32768, "kv_policy": "native installed defaults, not claimed equivalent to Strata int8", "configured_parallelism": 1, "has_vision_fixture": false }, { "candidate_id": "strata-iq2_xs", "backend": "strata", "model_id": "IQ2_XS", "frozen_request_model_name": "qwen3.8-flash-next-iq2_xs", "kind": "text", "manifest_digest": null, "manifest_label": null, "installed_size_bytes": null, "parameter_size": null, "quantization": "IQ2_XS", "configured_context_tokens": 32768, "kv_policy": "int8", "configured_parallelism": 1, "has_vision_fixture": false }, { "candidate_id": "strata-iq3_xxs", "backend": "strata", "model_id": "IQ3_XXS", "frozen_request_model_name": "qwen3.8-flash-next-iq3_xxs", "kind": "text", "manifest_digest": null, "manifest_label": null, "installed_size_bytes": null, "parameter_size": null, "quantization": "IQ3_XXS", "configured_context_tokens": 32768, "kv_policy": "int8", "configured_parallelism": 1, "has_vision_fixture": false }, { "candidate_id": "strata-iq3_s", "backend": "strata", "model_id": "IQ3_S", "frozen_request_model_name": "qwen3.8-flash-next-iq3_s", "kind": "text", "manifest_digest": null, "manifest_label": null, "installed_size_bytes": null, "parameter_size": null, "quantization": "IQ3_S", "configured_context_tokens": 32768, "kv_policy": "int8", "configured_parallelism": 1, "has_vision_fixture": false }, { "candidate_id": "strata-ud-q4_k_xl", "backend": "strata", "model_id": "UD-Q4_K_XL", "frozen_request_model_name": "qwen3.8-flash-next-unsloth-ud-q4_k_xl", "kind": "text", "manifest_digest": null, "manifest_label": null, "installed_size_bytes": null, "parameter_size": null, "quantization": "UD-Q4_K_XL", "configured_context_tokens": 32768, "kv_policy": "int8", "configured_parallelism": 1, "has_vision_fixture": false } ], "request_control_groups": [ { "candidate_id": "ollama-gptoss-ggml", "track": "code", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": "high" }, { "candidate_id": "ollama-gptoss-ggml", "track": "context", "planned_requests": 2, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": "low" }, { "candidate_id": "ollama-gptoss-ggml", "track": "performance", "planned_requests": 3, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": "low" }, { "candidate_id": "ollama-gptoss-ggml", "track": "qa", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": "high" }, { "candidate_id": "ollama-gptoss-ggml", "track": "workflow", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "native_think": "high" }, { "candidate_id": "ollama-gptoss-llamacpp", "track": "code", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": "high" }, { "candidate_id": "ollama-gptoss-llamacpp", "track": "context", "planned_requests": 2, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": "low" }, { "candidate_id": "ollama-gptoss-llamacpp", "track": "performance", "planned_requests": 3, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": "low" }, { "candidate_id": "ollama-gptoss-llamacpp", "track": "qa", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": "high" }, { "candidate_id": "ollama-gptoss-llamacpp", "track": "workflow", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "native_think": "high" }, { "candidate_id": "ollama-qwen35-ggml", "track": "code", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": true }, { "candidate_id": "ollama-qwen35-ggml", "track": "context", "planned_requests": 2, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": false }, { "candidate_id": "ollama-qwen35-ggml", "track": "performance", "planned_requests": 3, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": false }, { "candidate_id": "ollama-qwen35-ggml", "track": "qa", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": true }, { "candidate_id": "ollama-qwen35-ggml", "track": "vision", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": false }, { "candidate_id": "ollama-qwen35-ggml", "track": "vision-performance", "planned_requests": 3, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": false }, { "candidate_id": "ollama-qwen35-ggml", "track": "workflow", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "native_think": true }, { "candidate_id": "ollama-qwen35-llamacpp", "track": "code", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": true }, { "candidate_id": "ollama-qwen35-llamacpp", "track": "context", "planned_requests": 2, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": false }, { "candidate_id": "ollama-qwen35-llamacpp", "track": "performance", "planned_requests": 3, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": false }, { "candidate_id": "ollama-qwen35-llamacpp", "track": "qa", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": true }, { "candidate_id": "ollama-qwen35-llamacpp", "track": "vision", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": false }, { "candidate_id": "ollama-qwen35-llamacpp", "track": "vision-performance", "planned_requests": 3, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": false }, { "candidate_id": "ollama-qwen35-llamacpp", "track": "workflow", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "native_think": true }, { "candidate_id": "ollama-qwen3vl", "track": "code", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": false }, { "candidate_id": "ollama-qwen3vl", "track": "context", "planned_requests": 2, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": false }, { "candidate_id": "ollama-qwen3vl", "track": "performance", "planned_requests": 3, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": false }, { "candidate_id": "ollama-qwen3vl", "track": "qa", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": false }, { "candidate_id": "ollama-qwen3vl", "track": "vision", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": false }, { "candidate_id": "ollama-qwen3vl", "track": "vision-performance", "planned_requests": 3, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": false }, { "candidate_id": "ollama-qwen3vl", "track": "workflow", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "native_think": false }, { "candidate_id": "ollama-embedding", "track": "embedding-latency", "planned_requests": 3, "temperature": null, "top_p": null, "seed": null, "max_generated_tokens_requested": null, "reasoning_effort_requested": null, "reasoning_budget_tokens_requested": null, "native_think": null }, { "candidate_id": "ollama-embedding", "track": "embedding-quality", "planned_requests": 2, "temperature": null, "top_p": null, "seed": null, "max_generated_tokens_requested": null, "reasoning_effort_requested": null, "reasoning_budget_tokens_requested": null, "native_think": null }, { "candidate_id": "strata-iq2_xs", "track": "code", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": null }, { "candidate_id": "strata-iq2_xs", "track": "context", "planned_requests": 2, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": null }, { "candidate_id": "strata-iq2_xs", "track": "performance", "planned_requests": 3, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": null }, { "candidate_id": "strata-iq2_xs", "track": "qa", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": null }, { "candidate_id": "strata-iq2_xs", "track": "workflow", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "native_think": null }, { "candidate_id": "strata-iq3_xxs", "track": "code", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": null }, { "candidate_id": "strata-iq3_xxs", "track": "context", "planned_requests": 2, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": null }, { "candidate_id": "strata-iq3_xxs", "track": "performance", "planned_requests": 3, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": null }, { "candidate_id": "strata-iq3_xxs", "track": "qa", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": null }, { "candidate_id": "strata-iq3_xxs", "track": "workflow", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "native_think": null }, { "candidate_id": "strata-iq3_s", "track": "code", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": null }, { "candidate_id": "strata-iq3_s", "track": "context", "planned_requests": 2, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": null }, { "candidate_id": "strata-iq3_s", "track": "performance", "planned_requests": 3, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": null }, { "candidate_id": "strata-iq3_s", "track": "qa", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": null }, { "candidate_id": "strata-iq3_s", "track": "workflow", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "native_think": null }, { "candidate_id": "strata-ud-q4_k_xl", "track": "code", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": null }, { "candidate_id": "strata-ud-q4_k_xl", "track": "context", "planned_requests": 2, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": null }, { "candidate_id": "strata-ud-q4_k_xl", "track": "performance", "planned_requests": 3, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "native_think": null }, { "candidate_id": "strata-ud-q4_k_xl", "track": "qa", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "native_think": null }, { "candidate_id": "strata-ud-q4_k_xl", "track": "workflow", "planned_requests": 4, "temperature": 0, "top_p": 1, "seed": 41008, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "native_think": null } ], "cases": [ { "case_id": "higher_ring_buffer_minimal_fix", "track": "code" }, { "case_id": "higher_gpu_interval_percentile", "track": "code" }, { "case_id": "higher_safe_proposal_validator", "track": "code" }, { "case_id": "higher_bilingual_capture_status", "track": "code" }, { "case_id": "Q1-Chinese-conversation", "track": "qa" }, { "case_id": "Q2-Unpaired-percentiles", "track": "qa" }, { "case_id": "Q3-Overlapping-GPU-scopes", "track": "qa" }, { "case_id": "Q4-Untrusted-version-advice", "track": "qa" }, { "case_id": "context_lookup_6k_chars", "track": "context" }, { "case_id": "context_lookup_24k_chars", "track": "context" }, { "case_id": "workflow_no_pie_profile", "track": "workflow" }, { "case_id": "workflow_untrusted_excerpt", "track": "workflow" }, { "case_id": "workflow_read_only_capture", "track": "workflow" }, { "case_id": "workflow_destructive_approval", "track": "workflow" }, { "case_id": "speed_graphics_queue_512", "track": "performance" }, { "case_id": "vision_oblique_bleached_wood", "track": "vision" }, { "case_id": "vision_missing_wood_material", "track": "vision" }, { "case_id": "vision_catalog_material_smoke", "track": "vision" }, { "case_id": "vision_weathered_wood_fixed_exposure", "track": "vision" }, { "case_id": "vision_oblique_bleached_wood", "track": "vision-performance" }, { "case_id": "embedding_documents", "track": "embedding-quality" }, { "case_id": "embedding_queries", "track": "embedding-quality" }, { "case_id": "embedding_latency", "track": "embedding-latency" } ], "candidate_scoped_summaries": [ { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "status": "completed", "planned_scope_targets": 17, "attempted_or_intent_targets": 17, "response_records": 17, "observed_complete_response_deliveries": 6, "formal_complete_response_deliveries": 6, "formal_metrics_eligible": true, "code_checks": { "planned_cases": 4, "whole_cases_passed": 0, "planned_checks": 178, "executed_checks": 0, "passed_executed_checks": 0, "failed_executed_checks": 0, "not_executed_checks": 178 }, "tracks": { "context": { "planned": 2, "response_records": 2, "diagnostic_fixture_passes": 2, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "workflow": { "planned": 4, "response_records": 4, "diagnostic_fixture_passes": 2, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "performance": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" }, "vision": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": true, "group_status": "not_planned" }, "vision-performance": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" }, "embedding-latency": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" } }, "code_check_scope": "Own qualified segment only", "qa_grades_imported": false }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "status": "completed", "planned_scope_targets": 17, "attempted_or_intent_targets": 17, "response_records": 17, "observed_complete_response_deliveries": 6, "formal_complete_response_deliveries": 6, "formal_metrics_eligible": true, "code_checks": { "planned_cases": 4, "whole_cases_passed": 0, "planned_checks": 178, "executed_checks": 0, "passed_executed_checks": 0, "failed_executed_checks": 0, "not_executed_checks": 178 }, "tracks": { "context": { "planned": 2, "response_records": 2, "diagnostic_fixture_passes": 2, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "workflow": { "planned": 4, "response_records": 4, "diagnostic_fixture_passes": 2, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "performance": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" }, "vision": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": true, "group_status": "not_planned" }, "vision-performance": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" }, "embedding-latency": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" } }, "code_check_scope": "Own qualified segment only", "qa_grades_imported": false }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "status": "completed", "planned_scope_targets": 24, "attempted_or_intent_targets": 24, "response_records": 24, "observed_complete_response_deliveries": 19, "formal_complete_response_deliveries": 19, "formal_metrics_eligible": true, "code_checks": { "planned_cases": 4, "whole_cases_passed": 1, "planned_checks": 178, "executed_checks": 70, "passed_executed_checks": 63, "failed_executed_checks": 7, "not_executed_checks": 108 }, "tracks": { "context": { "planned": 2, "response_records": 2, "diagnostic_fixture_passes": 2, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "workflow": { "planned": 4, "response_records": 4, "diagnostic_fixture_passes": 1, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "performance": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" }, "vision": { "planned": 4, "response_records": 4, "diagnostic_fixture_passes": 3, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "vision-performance": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" }, "embedding-latency": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" } }, "code_check_scope": "Own qualified segment only", "qa_grades_imported": false }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "status": "completed", "planned_scope_targets": 24, "attempted_or_intent_targets": 24, "response_records": 24, "observed_complete_response_deliveries": 19, "formal_complete_response_deliveries": 19, "formal_metrics_eligible": true, "code_checks": { "planned_cases": 4, "whole_cases_passed": 1, "planned_checks": 178, "executed_checks": 70, "passed_executed_checks": 63, "failed_executed_checks": 7, "not_executed_checks": 108 }, "tracks": { "context": { "planned": 2, "response_records": 2, "diagnostic_fixture_passes": 2, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "workflow": { "planned": 4, "response_records": 4, "diagnostic_fixture_passes": 1, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "performance": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" }, "vision": { "planned": 4, "response_records": 4, "diagnostic_fixture_passes": 3, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "vision-performance": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" }, "embedding-latency": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" } }, "code_check_scope": "Own qualified segment only", "qa_grades_imported": false }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "status": "completed", "planned_scope_targets": 24, "attempted_or_intent_targets": 24, "response_records": 24, "observed_complete_response_deliveries": 20, "formal_complete_response_deliveries": 20, "formal_metrics_eligible": true, "code_checks": { "planned_cases": 4, "whole_cases_passed": 0, "planned_checks": 178, "executed_checks": 124, "passed_executed_checks": 79, "failed_executed_checks": 45, "not_executed_checks": 54 }, "tracks": { "context": { "planned": 2, "response_records": 2, "diagnostic_fixture_passes": 2, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "workflow": { "planned": 4, "response_records": 4, "diagnostic_fixture_passes": 2, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "performance": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" }, "vision": { "planned": 4, "response_records": 4, "diagnostic_fixture_passes": 3, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "vision-performance": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" }, "embedding-latency": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" } }, "code_check_scope": "Own qualified segment only", "qa_grades_imported": false }, { "source_study": "original-v3", "candidate_id": "ollama-embedding", "status": "completed", "planned_scope_targets": 5, "attempted_or_intent_targets": 5, "response_records": 5, "observed_complete_response_deliveries": 5, "formal_complete_response_deliveries": 5, "formal_metrics_eligible": true, "code_checks": { "planned_cases": 0, "whole_cases_passed": 0, "planned_checks": 0, "executed_checks": 0, "passed_executed_checks": 0, "failed_executed_checks": 0, "not_executed_checks": 0 }, "tracks": { "context": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": true, "group_status": "not_planned" }, "workflow": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": true, "group_status": "not_planned" }, "performance": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" }, "vision": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": true, "group_status": "not_planned" }, "vision-performance": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" }, "embedding-latency": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" } }, "code_check_scope": "Own qualified segment only", "qa_grades_imported": false }, { "source_study": "original-v3", "candidate_id": "strata-iq2_xs", "status": "failed", "planned_scope_targets": 17, "attempted_or_intent_targets": 3, "response_records": 3, "observed_complete_response_deliveries": 2, "formal_complete_response_deliveries": 0, "formal_metrics_eligible": false, "code_checks": { "planned_cases": 4, "whole_cases_passed": 2, "planned_checks": 178, "executed_checks": 70, "passed_executed_checks": 70, "failed_executed_checks": 0, "not_executed_checks": 108 }, "tracks": { "context": { "planned": 2, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": false, "group_status": "scope_observations_only" }, "workflow": { "planned": 4, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": false, "group_status": "scope_observations_only" }, "performance": { "planned": 3, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": false, "group_status": "scope_observations_only" }, "vision": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": false, "group_status": "not_planned" }, "vision-performance": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": false, "group_status": "not_planned" }, "embedding-latency": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": false, "group_status": "not_planned" } }, "code_check_scope": "Own-segment raw diagnostics only; no candidate pooling", "qa_grades_imported": false }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "status": "completed", "planned_scope_targets": 17, "attempted_or_intent_targets": 17, "response_records": 17, "observed_complete_response_deliveries": 14, "formal_complete_response_deliveries": 14, "formal_metrics_eligible": true, "code_checks": { "planned_cases": 4, "whole_cases_passed": 3, "planned_checks": 178, "executed_checks": 124, "passed_executed_checks": 124, "failed_executed_checks": 0, "not_executed_checks": 54 }, "tracks": { "context": { "planned": 2, "response_records": 2, "diagnostic_fixture_passes": 2, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "workflow": { "planned": 4, "response_records": 4, "diagnostic_fixture_passes": 1, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "performance": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" }, "vision": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": true, "group_status": "not_planned" }, "vision-performance": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" }, "embedding-latency": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" } }, "code_check_scope": "Own qualified segment only", "qa_grades_imported": false }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "status": "completed", "planned_scope_targets": 17, "attempted_or_intent_targets": 17, "response_records": 17, "observed_complete_response_deliveries": 14, "formal_complete_response_deliveries": 14, "formal_metrics_eligible": true, "code_checks": { "planned_cases": 4, "whole_cases_passed": 4, "planned_checks": 178, "executed_checks": 178, "passed_executed_checks": 178, "failed_executed_checks": 0, "not_executed_checks": 0 }, "tracks": { "context": { "planned": 2, "response_records": 2, "diagnostic_fixture_passes": 2, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "workflow": { "planned": 4, "response_records": 4, "diagnostic_fixture_passes": 1, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "performance": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" }, "vision": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": true, "group_status": "not_planned" }, "vision-performance": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" }, "embedding-latency": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" } }, "code_check_scope": "Own qualified segment only", "qa_grades_imported": false }, { "source_study": "original-v3", "candidate_id": "strata-ud-q4_k_xl", "status": "failed", "planned_scope_targets": 17, "attempted_or_intent_targets": 8, "response_records": 8, "observed_complete_response_deliveries": 7, "formal_complete_response_deliveries": 0, "formal_metrics_eligible": false, "code_checks": { "planned_cases": 4, "whole_cases_passed": 3, "planned_checks": 178, "executed_checks": 131, "passed_executed_checks": 131, "failed_executed_checks": 0, "not_executed_checks": 47 }, "tracks": { "context": { "planned": 2, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": false, "group_status": "scope_observations_only" }, "workflow": { "planned": 4, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": false, "group_status": "scope_observations_only" }, "performance": { "planned": 3, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": false, "group_status": "scope_observations_only" }, "vision": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": false, "group_status": "not_planned" }, "vision-performance": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": false, "group_status": "not_planned" }, "embedding-latency": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": false, "group_status": "not_planned" } }, "code_check_scope": "Own-segment raw diagnostics only; no candidate pooling", "qa_grades_imported": false }, { "source_study": "iq2-fresh-tail4", "candidate_id": "strata-iq2_xs", "status": "completed", "planned_scope_targets": 4, "attempted_or_intent_targets": 4, "response_records": 4, "observed_complete_response_deliveries": 1, "formal_complete_response_deliveries": 1, "formal_metrics_eligible": true, "code_checks": { "planned_cases": 0, "whole_cases_passed": 0, "planned_checks": 0, "executed_checks": 0, "passed_executed_checks": 0, "failed_executed_checks": 0, "not_executed_checks": 0 }, "tracks": { "context": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": true, "group_status": "not_planned" }, "workflow": { "planned": 1, "response_records": 1, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "performance": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" }, "vision": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": true, "group_status": "not_planned" }, "vision-performance": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" }, "embedding-latency": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" } }, "code_check_scope": "Own qualified segment only", "qa_grades_imported": false }, { "source_study": "q4-tail9", "candidate_id": "strata-ud-q4_k_xl", "status": "completed", "planned_scope_targets": 9, "attempted_or_intent_targets": 9, "response_records": 9, "observed_complete_response_deliveries": 6, "formal_complete_response_deliveries": 6, "formal_metrics_eligible": true, "code_checks": { "planned_cases": 0, "whole_cases_passed": 0, "planned_checks": 0, "executed_checks": 0, "passed_executed_checks": 0, "failed_executed_checks": 0, "not_executed_checks": 0 }, "tracks": { "context": { "planned": 2, "response_records": 2, "diagnostic_fixture_passes": 2, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "workflow": { "planned": 4, "response_records": 4, "diagnostic_fixture_passes": 1, "formal_metrics_eligible": true, "group_status": "scope_observations_only" }, "performance": { "planned": 3, "response_records": 3, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "qualified_group" }, "vision": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": true, "group_status": "not_planned" }, "vision-performance": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" }, "embedding-latency": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": true, "group_status": "not_planned" } }, "code_check_scope": "Own qualified segment only", "qa_grades_imported": false }, { "source_study": "iq2-interrupted-missing14", "candidate_id": "strata-iq2_xs", "status": "interrupted_unfinalized", "planned_scope_targets": 14, "attempted_or_intent_targets": 10, "response_records": 9, "observed_complete_response_deliveries": 9, "formal_complete_response_deliveries": 0, "formal_metrics_eligible": false, "code_checks": { "planned_cases": 1, "whole_cases_passed": 1, "planned_checks": 54, "executed_checks": 54, "passed_executed_checks": 54, "failed_executed_checks": 0, "not_executed_checks": 0 }, "tracks": { "context": { "planned": 2, "response_records": 2, "diagnostic_fixture_passes": 2, "formal_metrics_eligible": false, "group_status": "diagnostic_only" }, "workflow": { "planned": 4, "response_records": 2, "diagnostic_fixture_passes": 1, "formal_metrics_eligible": false, "group_status": "diagnostic_only" }, "performance": { "planned": 3, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": false, "group_status": "scope_observations_only" }, "vision": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": 0, "formal_metrics_eligible": false, "group_status": "not_planned" }, "vision-performance": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": false, "group_status": "not_planned" }, "embedding-latency": { "planned": 0, "response_records": 0, "diagnostic_fixture_passes": null, "formal_metrics_eligible": false, "group_status": "not_planned" } }, "code_check_scope": "Own-segment raw diagnostics only; no candidate pooling", "qa_grades_imported": false, "recorded_summary_status": "running", "status_note": "The source summary retained running when interrupted; this frozen snapshot has no sealed terminal lifecycle." } ], "segments": [ { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "source_protocol_sha256": "f08978de16ec13de5c0ca5d0fe1c2aa2a9057da507603abdfd9816c9eedded35", "source_terminal_sha256": "d906250fa7deb71ad87f4782ce179a0589ee9764290af131dc8628b1619e83cc", "status": "completed", "recorded_termination_verified": true, "recorded_release_qualified": true, "historical_release_verified": null, "lifecycle_qualified": true, "formal_metrics_eligible": true, "residency_cold_load": { "wall_seconds": 6.41117109998595, "confirmed": true, "sample_count": 7, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 14775, "sampled_peak_system_ram_used_gib": 20.29779052734375, "sampled_min_available_phys_gib": 41.35776138305664, "sampled_min_available_commit_gib": 27.293190002441406 }, "whole_system_lifecycle_samples": { "sample_count": 701, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 14878, "sampled_peak_system_ram_used_gib": 23.473018646240234, "sampled_min_available_phys_gib": 38.182533264160156, "sampled_min_available_commit_gib": 24.25702667236328 }, "resource_gate_passed": true, "watchdog_passed": true, "watchdog_closed": true, "watchdog_thread_alive": false, "unload_confirmed": true, "owned_process_tree_stop_confirmed": true, "recovered_baseline_passed": true }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "source_protocol_sha256": "f08978de16ec13de5c0ca5d0fe1c2aa2a9057da507603abdfd9816c9eedded35", "source_terminal_sha256": "b6550cc8e6dd775a7e65ddfae0af874e81d3712d578a931bc21d18ce381b7bad", "status": "completed", "recorded_termination_verified": true, "recorded_release_qualified": true, "historical_release_verified": null, "lifecycle_qualified": true, "formal_metrics_eligible": true, "residency_cold_load": { "wall_seconds": 5.931217799981823, "confirmed": true, "sample_count": 6, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 13800, "sampled_peak_system_ram_used_gib": 20.329730987548828, "sampled_min_available_phys_gib": 41.32582092285156, "sampled_min_available_commit_gib": 28.703704833984375 }, "whole_system_lifecycle_samples": { "sample_count": 681, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 14903, "sampled_peak_system_ram_used_gib": 23.241714477539062, "sampled_min_available_phys_gib": 38.41383743286133, "sampled_min_available_commit_gib": 24.30600357055664 }, "resource_gate_passed": true, "watchdog_passed": true, "watchdog_closed": true, "watchdog_thread_alive": false, "unload_confirmed": true, "owned_process_tree_stop_confirmed": true, "recovered_baseline_passed": true }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "source_protocol_sha256": "f08978de16ec13de5c0ca5d0fe1c2aa2a9057da507603abdfd9816c9eedded35", "source_terminal_sha256": "27f7b57552fd77e6afe48a4fe35bfec548230d7861fb06f68db76b6b5d7f4584", "status": "completed", "recorded_termination_verified": true, "recorded_release_qualified": true, "historical_release_verified": null, "lifecycle_qualified": true, "formal_metrics_eligible": true, "residency_cold_load": { "wall_seconds": 4.43688920000568, "confirmed": true, "sample_count": 5, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 9104, "sampled_peak_system_ram_used_gib": 20.483047485351562, "sampled_min_available_phys_gib": 41.17250442504883, "sampled_min_available_commit_gib": 31.560546875 }, "whole_system_lifecycle_samples": { "sample_count": 645, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 9516, "sampled_peak_system_ram_used_gib": 27.650070190429688, "sampled_min_available_phys_gib": 34.0054817199707, "sampled_min_available_commit_gib": 23.792736053466797 }, "resource_gate_passed": true, "watchdog_passed": true, "watchdog_closed": true, "watchdog_thread_alive": false, "unload_confirmed": true, "owned_process_tree_stop_confirmed": true, "recovered_baseline_passed": true }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "source_protocol_sha256": "f08978de16ec13de5c0ca5d0fe1c2aa2a9057da507603abdfd9816c9eedded35", "source_terminal_sha256": "9a0669f2a0dd0c52705a43ce2bdb91344f2b3f3b3ebd2e8dce37951c3eb40e3a", "status": "completed", "recorded_termination_verified": true, "recorded_release_qualified": true, "historical_release_verified": null, "lifecycle_qualified": true, "formal_metrics_eligible": true, "residency_cold_load": { "wall_seconds": 4.21003240000573, "confirmed": true, "sample_count": 4, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 8046, "sampled_peak_system_ram_used_gib": 22.218219757080078, "sampled_min_available_phys_gib": 39.43733215332031, "sampled_min_available_commit_gib": 31.072734832763672 }, "whole_system_lifecycle_samples": { "sample_count": 647, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 9343, "sampled_peak_system_ram_used_gib": 28.190502166748047, "sampled_min_available_phys_gib": 33.465049743652344, "sampled_min_available_commit_gib": 23.49386215209961 }, "resource_gate_passed": true, "watchdog_passed": true, "watchdog_closed": true, "watchdog_thread_alive": false, "unload_confirmed": true, "owned_process_tree_stop_confirmed": true, "recovered_baseline_passed": true }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "source_protocol_sha256": "f08978de16ec13de5c0ca5d0fe1c2aa2a9057da507603abdfd9816c9eedded35", "source_terminal_sha256": "3e0f466426745c064f2e3604b1c76ab7685a23ee38ea8488b4f320eb5d86a7ff", "status": "completed", "recorded_termination_verified": true, "recorded_release_qualified": true, "historical_release_verified": null, "lifecycle_qualified": true, "formal_metrics_eligible": true, "residency_cold_load": { "wall_seconds": 3.6671678999846336, "confirmed": true, "sample_count": 4, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 12820, "sampled_peak_system_ram_used_gib": 22.503692626953125, "sampled_min_available_phys_gib": 39.151859283447266, "sampled_min_available_commit_gib": 26.495044708251953 }, "whole_system_lifecycle_samples": { "sample_count": 336, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 13628, "sampled_peak_system_ram_used_gib": 30.632144927978516, "sampled_min_available_phys_gib": 31.023406982421875, "sampled_min_available_commit_gib": 17.745010375976562 }, "resource_gate_passed": true, "watchdog_passed": true, "watchdog_closed": true, "watchdog_thread_alive": false, "unload_confirmed": true, "owned_process_tree_stop_confirmed": true, "recovered_baseline_passed": true }, { "source_study": "original-v3", "candidate_id": "ollama-embedding", "source_protocol_sha256": "f08978de16ec13de5c0ca5d0fe1c2aa2a9057da507603abdfd9816c9eedded35", "source_terminal_sha256": "2901568a3f20812d8193046eb4cc9baa023a81659923aba715b017b797e0142b", "status": "completed", "recorded_termination_verified": true, "recorded_release_qualified": true, "historical_release_verified": null, "lifecycle_qualified": true, "formal_metrics_eligible": true, "residency_cold_load": { "wall_seconds": 4.183626899990486, "confirmed": true, "sample_count": 4, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 7579, "sampled_peak_system_ram_used_gib": 22.671085357666016, "sampled_min_available_phys_gib": 38.984466552734375, "sampled_min_available_commit_gib": 31.585529327392578 }, "whole_system_lifecycle_samples": { "sample_count": 67, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 7660, "sampled_peak_system_ram_used_gib": 22.81979751586914, "sampled_min_available_phys_gib": 38.83575439453125, "sampled_min_available_commit_gib": 31.332141876220703 }, "resource_gate_passed": true, "watchdog_passed": true, "watchdog_closed": true, "watchdog_thread_alive": false, "unload_confirmed": true, "owned_process_tree_stop_confirmed": true, "recovered_baseline_passed": true }, { "source_study": "original-v3", "candidate_id": "strata-iq2_xs", "source_protocol_sha256": "f08978de16ec13de5c0ca5d0fe1c2aa2a9057da507603abdfd9816c9eedded35", "source_terminal_sha256": "66c62851c0a2bfdf721131041cabeefdcee3d9b1809288954043df3983145d39", "status": "failed", "recorded_termination_verified": true, "recorded_release_qualified": true, "historical_release_verified": null, "lifecycle_qualified": true, "formal_metrics_eligible": false, "residency_cold_load": { "wall_seconds": 22.69642870000098, "confirmed": true, "sample_count": 21, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 14729, "sampled_peak_system_ram_used_gib": 35.81097412109375, "sampled_min_available_phys_gib": 25.84457778930664, "sampled_min_available_commit_gib": 10.686019897460938 }, "whole_system_lifecycle_samples": { "sample_count": 688, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 15266, "sampled_peak_system_ram_used_gib": 37.941062927246094, "sampled_min_available_phys_gib": 23.714488983154297, "sampled_min_available_commit_gib": 7.478633880615234 }, "resource_gate_passed": true, "watchdog_passed": true, "watchdog_closed": true, "watchdog_thread_alive": false, "unload_confirmed": true, "owned_process_tree_stop_confirmed": true, "recovered_baseline_passed": true }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "source_protocol_sha256": "f08978de16ec13de5c0ca5d0fe1c2aa2a9057da507603abdfd9816c9eedded35", "source_terminal_sha256": "72b7787f8a501d9b3da055390fbc683a1144a52fc9f10b6d6aaaf8420117c31c", "status": "completed", "recorded_termination_verified": true, "recorded_release_qualified": true, "historical_release_verified": null, "lifecycle_qualified": true, "formal_metrics_eligible": true, "residency_cold_load": { "wall_seconds": 13.58691360001103, "confirmed": true, "sample_count": 13, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 14982, "sampled_peak_system_ram_used_gib": 32.387413024902344, "sampled_min_available_phys_gib": 29.268138885498047, "sampled_min_available_commit_gib": 9.284015655517578 }, "whole_system_lifecycle_samples": { "sample_count": 3314, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 15085, "sampled_peak_system_ram_used_gib": 33.95308303833008, "sampled_min_available_phys_gib": 27.702468872070312, "sampled_min_available_commit_gib": 7.452205657958984 }, "resource_gate_passed": true, "watchdog_passed": true, "watchdog_closed": true, "watchdog_thread_alive": false, "unload_confirmed": true, "owned_process_tree_stop_confirmed": true, "recovered_baseline_passed": true }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "source_protocol_sha256": "f08978de16ec13de5c0ca5d0fe1c2aa2a9057da507603abdfd9816c9eedded35", "source_terminal_sha256": "7bc6aeaa857413a04eb9ac56337a3c9453b4947fbb58123fd494d9e9fec1835f", "status": "completed", "recorded_termination_verified": true, "recorded_release_qualified": true, "historical_release_verified": null, "lifecycle_qualified": true, "formal_metrics_eligible": true, "residency_cold_load": { "wall_seconds": 13.381607899995288, "confirmed": true, "sample_count": 12, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 14541, "sampled_peak_system_ram_used_gib": 32.09462356567383, "sampled_min_available_phys_gib": 29.560928344726562, "sampled_min_available_commit_gib": 9.751522064208984 }, "whole_system_lifecycle_samples": { "sample_count": 3607, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 15579, "sampled_peak_system_ram_used_gib": 35.767234802246094, "sampled_min_available_phys_gib": 25.888317108154297, "sampled_min_available_commit_gib": 4.744968414306641 }, "resource_gate_passed": true, "watchdog_passed": true, "watchdog_closed": true, "watchdog_thread_alive": false, "unload_confirmed": true, "owned_process_tree_stop_confirmed": true, "recovered_baseline_passed": true }, { "source_study": "original-v3", "candidate_id": "strata-ud-q4_k_xl", "source_protocol_sha256": "f08978de16ec13de5c0ca5d0fe1c2aa2a9057da507603abdfd9816c9eedded35", "source_terminal_sha256": "946ab21541a510acfa3fe4c0095db6e300bc5aecea93e5dbc394abc2a2884873", "status": "failed", "recorded_termination_verified": true, "recorded_release_qualified": false, "historical_release_verified": null, "lifecycle_qualified": false, "formal_metrics_eligible": false, "residency_cold_load": { "wall_seconds": 13.862549800018314, "confirmed": true, "sample_count": 13, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 14588, "sampled_peak_system_ram_used_gib": 31.452903747558594, "sampled_min_available_phys_gib": 30.202648162841797, "sampled_min_available_commit_gib": 10.194778442382812 }, "whole_system_lifecycle_samples": { "sample_count": 7981, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 15699, "sampled_peak_system_ram_used_gib": 33.47027587890625, "sampled_min_available_phys_gib": 28.18527603149414, "sampled_min_available_commit_gib": 6.6540069580078125 }, "resource_gate_passed": true, "watchdog_passed": false, "watchdog_closed": true, "watchdog_thread_alive": false, "unload_confirmed": null, "owned_process_tree_stop_confirmed": true, "recovered_baseline_passed": true }, { "source_study": "iq2-fresh-tail4", "candidate_id": "strata-iq2_xs", "source_protocol_sha256": "66e80c3f99e3d7df21c01c2a888cb4566a986152e61f66c5cb62fb6be7f95b5a", "source_terminal_sha256": "ec85f7dfbc9c6efec1d85fb032890fcffb279d2f8dde88fe37b0fc78ecc0408b", "status": "completed", "recorded_termination_verified": true, "recorded_release_qualified": true, "historical_release_verified": null, "lifecycle_qualified": true, "formal_metrics_eligible": true, "residency_cold_load": { "wall_seconds": 13.207565899996553, "confirmed": true, "sample_count": 12, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 15208, "sampled_peak_system_ram_used_gib": 32.00518035888672, "sampled_min_available_phys_gib": 29.650371551513672, "sampled_min_available_commit_gib": 10.348026275634766 }, "whole_system_lifecycle_samples": { "sample_count": 173, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 15701, "sampled_peak_system_ram_used_gib": 34.0243034362793, "sampled_min_available_phys_gib": 27.631248474121094, "sampled_min_available_commit_gib": 7.818332672119141 }, "resource_gate_passed": true, "watchdog_passed": true, "watchdog_closed": true, "watchdog_thread_alive": false, "unload_confirmed": true, "owned_process_tree_stop_confirmed": true, "recovered_baseline_passed": true }, { "source_study": "q4-tail9", "candidate_id": "strata-ud-q4_k_xl", "source_protocol_sha256": "6736c5d4b1aa246e50a7756d8342e6f370ec6aeaf1c88fcde537348e7da829c0", "source_terminal_sha256": "b3b9210022f3409412702ed090f768b0fbede5e433bfeb6e1c2a56cc2710bdb2", "status": "completed", "recorded_termination_verified": true, "recorded_release_qualified": true, "historical_release_verified": null, "lifecycle_qualified": true, "formal_metrics_eligible": true, "residency_cold_load": { "wall_seconds": 19.64574680000078, "confirmed": true, "sample_count": 18, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 15030, "sampled_peak_system_ram_used_gib": 37.535865783691406, "sampled_min_available_phys_gib": 24.119686126708984, "sampled_min_available_commit_gib": 9.78656005859375 }, "whole_system_lifecycle_samples": { "sample_count": 1349, "sampling_error_count": 0, "sampled_peak_gpu_used_mib": 15720, "sampled_peak_system_ram_used_gib": 39.83585739135742, "sampled_min_available_phys_gib": 21.81969451904297, "sampled_min_available_commit_gib": 2.416065216064453 }, "resource_gate_passed": true, "watchdog_passed": true, "watchdog_closed": true, "watchdog_thread_alive": false, "unload_confirmed": true, "owned_process_tree_stop_confirmed": true, "recovered_baseline_passed": true }, { "source_study": "iq2-interrupted-missing14", "candidate_id": "strata-iq2_xs", "source_protocol_sha256": "a3f2e09ea12183afafc261f424a8925f0ec2a3ac407d45f42a7b34beb458e543", "source_terminal_sha256": null, "status": "interrupted_unfinalized", "recorded_termination_verified": false, "recorded_release_qualified": null, "historical_release_verified": false, "lifecycle_qualified": false, "formal_metrics_eligible": false, "residency_cold_load": { "wall_seconds": null, "confirmed": null, "sample_count": null, "sampling_error_count": null, "sampled_peak_gpu_used_mib": null, "sampled_peak_system_ram_used_gib": null, "sampled_min_available_phys_gib": null, "sampled_min_available_commit_gib": null }, "whole_system_lifecycle_samples": null, "resource_gate_passed": null, "watchdog_passed": null, "watchdog_closed": null, "watchdog_thread_alive": null, "unload_confirmed": null, "owned_process_tree_stop_confirmed": null, "recovered_baseline_passed": null } ], "performance_groups": [ { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "track": "performance", "case_id": "speed_graphics_queue_512", "source_terminal_sha256": "d906250fa7deb71ad87f4782ce179a0589ee9764290af131dc8628b1619e83cc", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 3, "median": 172.42511546926067, "min": 172.28476518124896, "max": 172.89639622451003 }, "api_prompt_tokens_per_second": { "n": 3, "median": 21774.588158538576, "min": 5207.622242593279, "max": 28069.806560134566 }, "wall_seconds": { "n": 3, "median": 3.023247699980857, "min": 2.998282199987443, "max": 3.071733300021151 }, "ttft_any_seconds": { "n": 3, "median": 0.07289459998719394, "min": 0.05758589998004027, "max": 0.12060290001682006 }, "ttft_answer_seconds": { "n": 3, "median": 0.1416540999780409, "min": 0.12600279998150654, "max": 0.19186330001684837 }, "whole_system_ram_used_gib": { "n": 3, "median": 23.465232849121094, "min": 23.46326446533203, "max": 23.47713851928711 }, "whole_system_gpu_used_mib": { "n": 3, "median": 14798, "min": 14789, "max": 14802 }, "api_decode_three_sample_measured": true }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "track": "performance", "case_id": "speed_graphics_queue_512", "source_terminal_sha256": "b6550cc8e6dd775a7e65ddfae0af874e81d3712d578a931bc21d18ce381b7bad", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 3, "median": 179.09249077778608, "min": 178.7301987361122, "max": 179.24440780203295 }, "api_prompt_tokens_per_second": { "n": 3, "median": 22225.92191792225, "min": 5419.889166311431, "max": 27946.40988067825 }, "wall_seconds": { "n": 3, "median": 2.9008435000141617, "min": 2.88999560000957, "max": 2.9347764999838546 }, "ttft_any_seconds": { "n": 3, "median": 0.05597010001656599, "min": 0.05173020000802353, "max": 0.09583599999314174 }, "ttft_answer_seconds": { "n": 3, "median": 0.12228010001126677, "min": 0.11897830001544207, "max": 0.16333330000634305 }, "whole_system_ram_used_gib": { "n": 3, "median": 23.218303680419922, "min": 23.195819854736328, "max": 23.245346069335938 }, "whole_system_gpu_used_mib": { "n": 3, "median": 14791, "min": 14790, "max": 14791 }, "api_decode_three_sample_measured": true }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "track": "performance", "case_id": "speed_graphics_queue_512", "source_terminal_sha256": "27f7b57552fd77e6afe48a4fe35bfec548230d7861fb06f68db76b6b5d7f4584", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 3, "median": 127.14523437560142, "min": 126.08564170203796, "max": 128.30273029212427 }, "api_prompt_tokens_per_second": { "n": 3, "median": 7332.595462091865, "min": 3399.778693651074, "max": 7634.137558516384 }, "wall_seconds": { "n": 3, "median": 4.126083200011635, "min": 4.0917132000031415, "max": 4.130043699988164 }, "ttft_any_seconds": { "n": 3, "median": 0.06527240001014434, "min": 0.06464139997842722, "max": 0.13830089999828488 }, "ttft_answer_seconds": { "n": 3, "median": 0.06527240001014434, "min": 0.06464139997842722, "max": 0.13830089999828488 }, "whole_system_ram_used_gib": { "n": 3, "median": 27.357128143310547, "min": 27.264530181884766, "max": 27.36655044555664 }, "whole_system_gpu_used_mib": { "n": 3, "median": 9403, "min": 9391, "max": 9414 }, "api_decode_three_sample_measured": true }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "track": "vision-performance", "case_id": "vision_oblique_bleached_wood", "source_terminal_sha256": "27f7b57552fd77e6afe48a4fe35bfec548230d7861fb06f68db76b6b5d7f4584", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 3, "median": 126.46611518793189, "min": 125.09321850417378, "max": 127.54664037243619 }, "api_prompt_tokens_per_second": { "n": 3, "median": 150967.56326375186, "min": 144389.6301691885, "max": 158311.24986386902 }, "wall_seconds": { "n": 3, "median": 0.7387351999932434, "min": 0.7119180999870878, "max": 0.7759215999976732 }, "ttft_any_seconds": { "n": 3, "median": 0.12193330001900904, "min": 0.10027649998664856, "max": 0.15225359998294152 }, "ttft_answer_seconds": { "n": 3, "median": 0.12193330001900904, "min": 0.10027649998664856, "max": 0.15225359998294152 }, "whole_system_ram_used_gib": { "n": 3, "median": 27.506149291992188, "min": 27.480133056640625, "max": 27.70001983642578 }, "whole_system_gpu_used_mib": { "n": 3, "median": 9512, "min": 9506, "max": 9512 }, "api_decode_three_sample_measured": true }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "track": "performance", "case_id": "speed_graphics_queue_512", "source_terminal_sha256": "9a0669f2a0dd0c52705a43ce2bdb91344f2b3f3b3ebd2e8dce37951c3eb40e3a", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 3, "median": 130.3620959984438, "min": 130.22300435153394, "max": 130.46932773130658 }, "api_prompt_tokens_per_second": { "n": 3, "median": 7199.37514857201, "min": 3328.7275467906043, "max": 7658.1295379836 }, "wall_seconds": { "n": 3, "median": 4.00571990001481, "min": 3.9923479999997653, "max": 4.074568099982571 }, "ttft_any_seconds": { "n": 3, "median": 0.08152370000607334, "min": 0.060955500026466325, "max": 0.14696710000862367 }, "ttft_answer_seconds": { "n": 3, "median": 0.08152370000607334, "min": 0.060955500026466325, "max": 0.14696710000862367 }, "whole_system_ram_used_gib": { "n": 3, "median": 28.12236785888672, "min": 28.075103759765625, "max": 28.138076782226562 }, "whole_system_gpu_used_mib": { "n": 3, "median": 9333, "min": 9333, "max": 9343 }, "api_decode_three_sample_measured": true }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "track": "vision-performance", "case_id": "vision_oblique_bleached_wood", "source_terminal_sha256": "9a0669f2a0dd0c52705a43ce2bdb91344f2b3f3b3ebd2e8dce37951c3eb40e3a", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 3, "median": 129.10722354915757, "min": 129.03930239367907, "max": 129.4820717131474 }, "api_prompt_tokens_per_second": { "n": 3, "median": 153475.2771423544, "min": 124899.75942261427, "max": 156313.84637442202 }, "wall_seconds": { "n": 3, "median": 0.721753699996043, "min": 0.7017243999871425, "max": 0.7385091000178363 }, "ttft_any_seconds": { "n": 3, "median": 0.11939439998241141, "min": 0.09758039997541346, "max": 0.13414450001437217 }, "ttft_answer_seconds": { "n": 3, "median": 0.11939439998241141, "min": 0.09758039997541346, "max": 0.13414450001437217 }, "whole_system_ram_used_gib": { "n": 3, "median": 27.889907836914062, "min": 27.859729766845703, "max": 28.119911193847656 }, "whole_system_gpu_used_mib": { "n": 3, "median": 9333, "min": 9333, "max": 9333 }, "api_decode_three_sample_measured": true }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "track": "performance", "case_id": "speed_graphics_queue_512", "source_terminal_sha256": "3e0f466426745c064f2e3604b1c76ab7685a23ee38ea8488b4f320eb5d86a7ff", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 3, "median": 150.63019023239707, "min": 150.16463851529406, "max": 150.70467681941312 }, "api_prompt_tokens_per_second": { "n": 3, "median": 18689.909246113755, "min": 9718.71787683394, "max": 18809.911376379092 }, "wall_seconds": { "n": 3, "median": 3.446389100019587, "min": 3.4448866000166163, "max": 3.5724170000175945 }, "ttft_any_seconds": { "n": 3, "median": 0.04911980000906624, "min": 0.046404799999436364, "max": 0.1630795999953989 }, "ttft_answer_seconds": { "n": 3, "median": 0.04911980000906624, "min": 0.046404799999436364, "max": 0.1630795999953989 }, "whole_system_ram_used_gib": { "n": 3, "median": 30.557086944580078, "min": 30.485523223876953, "max": 30.591636657714844 }, "whole_system_gpu_used_mib": { "n": 3, "median": 13466, "min": 13466, "max": 13466 }, "api_decode_three_sample_measured": true }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "track": "vision-performance", "case_id": "vision_oblique_bleached_wood", "source_terminal_sha256": "3e0f466426745c064f2e3604b1c76ab7685a23ee38ea8488b4f320eb5d86a7ff", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 3, "median": 136.46066521325227, "min": 132.36726242440096, "max": 136.46908975117137 }, "api_prompt_tokens_per_second": { "n": 3, "median": 525750, "min": 356652.25133553805, "max": 526473.9016147202 }, "wall_seconds": { "n": 3, "median": 0.7311061000218615, "min": 0.7205277000030037, "max": 0.7951121999940369 }, "ttft_any_seconds": { "n": 3, "median": 0.11522370000602677, "min": 0.08513699998729862, "max": 0.17972780001582578 }, "ttft_answer_seconds": { "n": 3, "median": 0.11522370000602677, "min": 0.08513699998729862, "max": 0.17972780001582578 }, "whole_system_ram_used_gib": { "n": 3, "median": 30.245220184326172, "min": 30.219398498535156, "max": 30.600994110107422 }, "whole_system_gpu_used_mib": { "n": 3, "median": 13526, "min": 13466, "max": 13627 }, "api_decode_three_sample_measured": true }, { "source_study": "original-v3", "candidate_id": "ollama-embedding", "track": "embedding-latency", "case_id": "embedding_latency", "source_terminal_sha256": "2901568a3f20812d8193046eb4cc9baa023a81659923aba715b017b797e0142b", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 0, "median": null, "min": null, "max": null }, "api_prompt_tokens_per_second": { "n": 0, "median": null, "min": null, "max": null }, "wall_seconds": { "n": 3, "median": 0.024272600014228374, "min": 0.02216799999587238, "max": 0.02772949999780394 }, "ttft_any_seconds": { "n": 0, "median": null, "min": null, "max": null }, "ttft_answer_seconds": { "n": 0, "median": null, "min": null, "max": null }, "whole_system_ram_used_gib": { "n": 3, "median": 22.504531860351562, "min": 22.494430541992188, "max": 22.506221771240234 }, "whole_system_gpu_used_mib": { "n": 3, "median": 7627, "min": 7627, "max": 7627 }, "api_decode_three_sample_measured": false }, { "source_study": "original-v3", "candidate_id": "strata-iq2_xs", "track": "performance", "case_id": "speed_graphics_queue_512", "source_terminal_sha256": "66c62851c0a2bfdf721131041cabeefdcee3d9b1809288954043df3983145d39", "planned_repetitions": 3, "observed_bounded_repetitions": 0, "diagnostic_observed_bounded_repetitions": 0, "formal_metrics_eligible": false, "qualified_three_sample_group": false, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 0, "median": null, "min": null, "max": null }, "api_prompt_tokens_per_second": { "n": 0, "median": null, "min": null, "max": null }, "wall_seconds": { "n": 0, "median": null, "min": null, "max": null }, "ttft_any_seconds": { "n": 0, "median": null, "min": null, "max": null }, "ttft_answer_seconds": { "n": 0, "median": null, "min": null, "max": null }, "whole_system_ram_used_gib": { "n": 0, "median": null, "min": null, "max": null }, "whole_system_gpu_used_mib": { "n": 0, "median": null, "min": null, "max": null }, "api_decode_three_sample_measured": false }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "track": "performance", "case_id": "speed_graphics_queue_512", "source_terminal_sha256": "72b7787f8a501d9b3da055390fbc683a1144a52fc9f10b6d6aaaf8420117c31c", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 3, "median": 12.7, "min": 10.3, "max": 12.9 }, "api_prompt_tokens_per_second": { "n": 3, "median": 11.9, "min": 11.3, "max": 46 }, "wall_seconds": { "n": 3, "median": 40.82330469999579, "min": 40.43938310001977, "max": 54.16855060000671 }, "ttft_any_seconds": { "n": 3, "median": 0.8674681000120472, "min": 0.8293482000008225, "max": 4.755726200004574 }, "ttft_answer_seconds": { "n": 3, "median": 0.8674681000120472, "min": 0.8293482000008225, "max": 4.755726200004574 }, "whole_system_ram_used_gib": { "n": 3, "median": 32.89594650268555, "min": 32.863922119140625, "max": 32.898563385009766 }, "whole_system_gpu_used_mib": { "n": 3, "median": 14885, "min": 14885, "max": 14885 }, "api_decode_three_sample_measured": true }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "track": "performance", "case_id": "speed_graphics_queue_512", "source_terminal_sha256": "7bc6aeaa857413a04eb9ac56337a3c9453b4947fbb58123fd494d9e9fec1835f", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 3, "median": 8.3, "min": 8.1, "max": 8.7 }, "api_prompt_tokens_per_second": { "n": 3, "median": 10, "min": 9.5, "max": 38.5 }, "wall_seconds": { "n": 3, "median": 64.28774090000661, "min": 59.470488499995554, "max": 67.45779540002695 }, "ttft_any_seconds": { "n": 3, "median": 0.9798636000195984, "min": 0.9598735999898054, "max": 5.720312800025567 }, "ttft_answer_seconds": { "n": 3, "median": 0.9798636000195984, "min": 0.9598735999898054, "max": 5.720312800025567 }, "whole_system_ram_used_gib": { "n": 3, "median": 35.163230895996094, "min": 35.07213592529297, "max": 35.319759368896484 }, "whole_system_gpu_used_mib": { "n": 3, "median": 15064, "min": 15058, "max": 15131 }, "api_decode_three_sample_measured": true }, { "source_study": "original-v3", "candidate_id": "strata-ud-q4_k_xl", "track": "performance", "case_id": "speed_graphics_queue_512", "source_terminal_sha256": "946ab21541a510acfa3fe4c0095db6e300bc5aecea93e5dbc394abc2a2884873", "planned_repetitions": 3, "observed_bounded_repetitions": 0, "diagnostic_observed_bounded_repetitions": 0, "formal_metrics_eligible": false, "qualified_three_sample_group": false, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 0, "median": null, "min": null, "max": null }, "api_prompt_tokens_per_second": { "n": 0, "median": null, "min": null, "max": null }, "wall_seconds": { "n": 0, "median": null, "min": null, "max": null }, "ttft_any_seconds": { "n": 0, "median": null, "min": null, "max": null }, "ttft_answer_seconds": { "n": 0, "median": null, "min": null, "max": null }, "whole_system_ram_used_gib": { "n": 0, "median": null, "min": null, "max": null }, "whole_system_gpu_used_mib": { "n": 0, "median": null, "min": null, "max": null }, "api_decode_three_sample_measured": false }, { "source_study": "iq2-fresh-tail4", "candidate_id": "strata-iq2_xs", "track": "performance", "case_id": "speed_graphics_queue_512", "source_terminal_sha256": "ec85f7dfbc9c6efec1d85fb032890fcffb279d2f8dde88fe37b0fc78ecc0408b", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 3, "median": 21.4, "min": 18.9, "max": 22 }, "api_prompt_tokens_per_second": { "n": 3, "median": 14.2, "min": 12.7, "max": 58.2 }, "wall_seconds": { "n": 3, "median": 24.462002799999027, "min": 23.851367800001753, "max": 30.813348399999086 }, "ttft_any_seconds": { "n": 3, "median": 0.7012241999982507, "min": 0.6500033999982406, "max": 3.73461920000409 }, "ttft_answer_seconds": { "n": 3, "median": 0.7012241999982507, "min": 0.6500033999982406, "max": 3.73461920000409 }, "whole_system_ram_used_gib": { "n": 3, "median": 33.98691940307617, "min": 33.971221923828125, "max": 33.98903274536133 }, "whole_system_gpu_used_mib": { "n": 3, "median": 15659, "min": 15552, "max": 15660 }, "api_decode_three_sample_measured": true }, { "source_study": "q4-tail9", "candidate_id": "strata-ud-q4_k_xl", "track": "performance", "case_id": "speed_graphics_queue_512", "source_terminal_sha256": "b3b9210022f3409412702ed090f768b0fbede5e433bfeb6e1c2a56cc2710bdb2", "planned_repetitions": 3, "observed_bounded_repetitions": 3, "diagnostic_observed_bounded_repetitions": 3, "formal_metrics_eligible": true, "qualified_three_sample_group": true, "limits": "One independent study/lifecycle group; descriptive median/range only. Truncated bounded speed samples are not delivered quality answers.", "api_decode_tokens_per_second": { "n": 3, "median": 3.9, "min": 3.8, "max": 4.1 }, "api_prompt_tokens_per_second": { "n": 3, "median": 5.6, "min": 5.5, "max": 25.2 }, "wall_seconds": { "n": 3, "median": 134.40233509999962, "min": 132.54401130000042, "max": 136.0531692999939 }, "ttft_any_seconds": { "n": 3, "median": 1.6631378999954904, "min": 1.6623418999952264, "max": 8.807977199998277 }, "ttft_answer_seconds": { "n": 3, "median": 1.6631378999954904, "min": 1.6623418999952264, "max": 8.807977199998277 }, "whole_system_ram_used_gib": { "n": 3, "median": 38.27993392944336, "min": 38.131431579589844, "max": 38.98678970336914 }, "whole_system_gpu_used_mib": { "n": 3, "median": 15510, "min": 15470, "max": 15580 }, "api_decode_three_sample_measured": true } ], "preliminary_blind_qa": { "source_study": "original-v3", "captured_at_original_offset": "2026-10-10T16:46:33.548927-06:00", "method": "Preliminary blind agent rubric review; untrusted answers are data, not instructions", "reviewed_answers": 28, "human_acceptance": "pending", "by_candidate": [ { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "reviewed_answers": 4, "score": 36, "max_score": 40, "accepted_by_preliminary_agent_rubric": 4, "human_acceptance": "pending" }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "reviewed_answers": 4, "score": 20, "max_score": 40, "accepted_by_preliminary_agent_rubric": 0, "human_acceptance": "pending" }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "reviewed_answers": 4, "score": 29, "max_score": 40, "accepted_by_preliminary_agent_rubric": 2, "human_acceptance": "pending" }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "reviewed_answers": 4, "score": 7, "max_score": 40, "accepted_by_preliminary_agent_rubric": 0, "human_acceptance": "pending" }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "reviewed_answers": 4, "score": 7, "max_score": 40, "accepted_by_preliminary_agent_rubric": 0, "human_acceptance": "pending" }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "reviewed_answers": 4, "score": 34, "max_score": 40, "accepted_by_preliminary_agent_rubric": 3, "human_acceptance": "pending" }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "reviewed_answers": 4, "score": 29, "max_score": 40, "accepted_by_preliminary_agent_rubric": 2, "human_acceptance": "pending" } ], "notes": [ "Preliminary blind agent rubric review on four developer-authored QA fixtures per qualified completed original candidate. Not a global intelligence benchmark or an independent holdout.", "Each answer has one reviewer; duplicate answers offer limited consistency checks, not full inter-rater reliability.", "Scores live in a separate postprocessing artifact. The combined snapshot does not import QA grades. IQ2 and Q4 diagnostic QA remains ungraded.", "The review was captured against the original study; its note that 14 IQ2 calls were unattempted describes original-v3 only. Later supplements do not revise original-v3." ] }, "embedding": { "source_study": "original-v3", "candidate_id": "ollama-embedding", "document_vectors": 24, "query_vectors": 8, "dimensions": 1024, "document_batch_wall_seconds": 8.380409200006397, "query_batch_wall_seconds": 0.14361640001880005, "retrieval": { "passed": true, "count": 8, "query_count": 8, "top1": 1, "top3": 1, "mrr": 1, "limits": "Eight developer-authored retrieval queries, not full Chroma/RAG or independent embedding acceptance." }, "query_results": [ { "query_id": "zh_tail_latency", "gold_ids": [ "frame_percentile" ], "top3_ids": [ "frame_percentile", "cpu_timeline", "gpu_overlap" ], "rank": 1, "top3_hit": true, "reciprocal_rank": 1 }, { "query_id": "en_preserve_shadows", "gold_ids": [ "shadow_baseline" ], "top3_ids": [ "shadow_baseline", "mesh_lod", "occlusion" ], "rank": 1, "top3_hit": true, "reciprocal_rank": 1 }, { "query_id": "zh_delete_approval", "gold_ids": [ "approval_guard" ], "top3_ids": [ "approval_guard", "undo_recovery", "light_units" ], "rank": 1, "top3_hit": true, "reciprocal_rank": 1 }, { "query_id": "en_trace_export", "gold_ids": [ "insights_scopes" ], "top3_ids": [ "insights_scopes", "cpu_timeline", "draw_submissions" ], "rank": 1, "top3_hit": true, "reciprocal_rank": 1 }, { "query_id": "zh_dark_material", "gold_ids": [ "material_appearance" ], "top3_ids": [ "material_appearance", "image_evidence", "texture_streaming" ], "rank": 1, "top3_hit": true, "reciprocal_rank": 1 }, { "query_id": "en_project_asset", "gold_ids": [ "asset_retrieval" ], "top3_ids": [ "asset_retrieval", "material_appearance", "nanite_geometry" ], "rank": 1, "top3_hit": true, "reciprocal_rank": 1 }, { "query_id": "zh_context_identity", "gold_ids": [ "context_budget" ], "top3_ids": [ "context_budget", "undo_recovery", "camera_units" ], "rank": 1, "top3_hit": true, "reciprocal_rank": 1 }, { "query_id": "en_injected_forum", "gold_ids": [ "source_trust" ], "top3_ids": [ "source_trust", "undo_recovery", "context_budget" ], "rank": 1, "top3_hit": true, "reciprocal_rank": 1 } ] }, "failure_and_recovery_notes": { "original_iq2": { "candidate_id": "strata-iq2_xs", "source_study": "original-v3", "request_id": "code--higher_safe_proposal_validator--0", "request_status": "transport_or_response_failure", "failure": "Malformed structured JSON output", "remaining_original_targets": 14, "status_remains": "failed", "diagnosis_limits": "Not proof of network failure, OOM or overall model inability." }, "original_q4": { "candidate_id": "strata-ud-q4_k_xl", "source_study": "original-v3", "failure": "ValueError: Owned guard failed: deadline", "status_remains": "failed", "formal_metrics_eligible": false }, "interrupted_iq2": { "candidate_id": "strata-iq2_xs", "source_study": "iq2-interrupted-missing14", "persisted_response_records": 9, "persisted_generation_intents": 10, "pending_request_id": "workflow--workflow_read_only_capture--0", "pending_request_sha256": "8f8871d95214ee7cee1867df074979a519601a86b87ac97f2e474e694e6e84ca", "pending_dispatch_verified": false, "pending_formal_response_recorded": false, "only_keepalive_comments": true, "keepalive_comment_count": 41, "lifecycle_qualified": false, "formal_metrics_eligible": false, "diagnostic_only": true }, "tail_scopes": [ "IQ2 fresh tail4 copies original requests[13:] and interrupted supplement requests[10:]. It excludes original failure, nine interrupted responses and pending intent.", "Q4 tail9 contains two context, four workflow and three 512-token speed requests. It does not repeat original QA/code or erase deadline failure.", "Tail lifecycles are independently qualified; no quality, cold-load, memory or timing pooling across studies." ] }, "source_artifacts": [ { "source_study": "installed-options-interrupted-evidence", "artifact": "snapshot.json", "sha256": "8dade99436f7fa2454c5ccfaca483b6853c4642ebb063b294937d84c0e74aac0", "bytes": 1547885, "expected_sha256": "8dade99436f7fa2454c5ccfaca483b6853c4642ebb063b294937d84c0e74aac0", "recorded_digest_matches": true }, { "source_study": "installed-options-interrupted-evidence", "artifact": "report.html", "sha256": "1b6e003d2c2dc139cc0812222d1101b97061fbbd64e963fc32b93b2b3b602d0b", "bytes": 474002, "expected_sha256": "1b6e003d2c2dc139cc0812222d1101b97061fbbd64e963fc32b93b2b3b602d0b", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "protocol.json", "sha256": "f08978de16ec13de5c0ca5d0fe1c2aa2a9057da507603abdfd9816c9eedded35", "bytes": 49638486, "expected_sha256": "f08978de16ec13de5c0ca5d0fe1c2aa2a9057da507603abdfd9816c9eedded35", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "summary.json", "sha256": "a2ca5b0b4a8204597bda6b88bb40bf41e373da6b2cb5bf0abbd93f048eededba", "bytes": 974187, "expected_sha256": "a2ca5b0b4a8204597bda6b88bb40bf41e373da6b2cb5bf0abbd93f048eededba", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "checkpoint.json", "sha256": "6896513cfb2efeebfbe945a447fac657d032a1be0c37d142823f8e42951219e2", "bytes": 11999, "expected_sha256": "6896513cfb2efeebfbe945a447fac657d032a1be0c37d142823f8e42951219e2", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "reviewed-final-snapshot-20261010T224700Z.review.json", "sha256": "2d29232ff6eea3f25c6f15693c729e0725358af48636e835d6883991f6c8c789", "bytes": 84581, "expected_sha256": null, "recorded_digest_matches": null }, { "source_study": "original-v3", "artifact": "reviewed-final-snapshot-20261010T224700Z.html", "sha256": "a41ac7c450ce951a85ff8d4e1364125fa0f5ed7e8a5e0a212a207d3d6c7ff3e9", "bytes": 5894540, "expected_sha256": null, "recorded_digest_matches": null }, { "source_study": "iq2-interrupted-missing14", "artifact": "protocol.json", "sha256": "a3f2e09ea12183afafc261f424a8925f0ec2a3ac407d45f42a7b34beb458e543", "bytes": 25040372, "expected_sha256": "a3f2e09ea12183afafc261f424a8925f0ec2a3ac407d45f42a7b34beb458e543", "recorded_digest_matches": true }, { "source_study": "iq2-interrupted-missing14", "artifact": "summary.json", "sha256": "a4f1176d832f6f90e0764585470932cc7f792408e3e23a587af9500bfb088146", "bytes": 59909, "expected_sha256": "a4f1176d832f6f90e0764585470932cc7f792408e3e23a587af9500bfb088146", "recorded_digest_matches": true }, { "source_study": "iq2-interrupted-observation", "artifact": "recovery-manifest.json", "sha256": "ee46f3976369c7598c27b1f5cfd1a7f1101a4933982f51f52b367a4d2b1e2cbe", "bytes": 84243, "expected_sha256": "ee46f3976369c7598c27b1f5cfd1a7f1101a4933982f51f52b367a4d2b1e2cbe", "recorded_digest_matches": true }, { "source_study": "iq2-fresh-tail4", "artifact": "protocol.json", "sha256": "66e80c3f99e3d7df21c01c2a888cb4566a986152e61f66c5cb62fb6be7f95b5a", "bytes": 24833306, "expected_sha256": "66e80c3f99e3d7df21c01c2a888cb4566a986152e61f66c5cb62fb6be7f95b5a", "recorded_digest_matches": true }, { "source_study": "iq2-fresh-tail4", "artifact": "report.html", "sha256": "334727e8cc43dc7f952ec96efaa53c4d30ba754cb89d3a2ca4d4419c5856c592", "bytes": 4140, "expected_sha256": null, "recorded_digest_matches": null }, { "source_study": "q4-tail9", "artifact": "protocol.json", "sha256": "6736c5d4b1aa246e50a7756d8342e6f370ec6aeaf1c88fcde537348e7da829c0", "bytes": 25000205, "expected_sha256": "6736c5d4b1aa246e50a7756d8342e6f370ec6aeaf1c88fcde537348e7da829c0", "recorded_digest_matches": true }, { "source_study": "q4-tail9", "artifact": "report.html", "sha256": "2da050b4b75d06abb5c360bc39b8174374543d373ac048c47858b9169211bff6", "bytes": 4964, "expected_sha256": null, "recorded_digest_matches": null }, { "source_study": "original-v3", "artifact": "ollama-gptoss-ggml/terminal.json", "sha256": "d906250fa7deb71ad87f4782ce179a0589ee9764290af131dc8628b1619e83cc", "bytes": 104249, "expected_sha256": "d906250fa7deb71ad87f4782ce179a0589ee9764290af131dc8628b1619e83cc", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "ollama-gptoss-llamacpp/terminal.json", "sha256": "b6550cc8e6dd775a7e65ddfae0af874e81d3712d578a931bc21d18ce381b7bad", "bytes": 105093, "expected_sha256": "b6550cc8e6dd775a7e65ddfae0af874e81d3712d578a931bc21d18ce381b7bad", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "ollama-qwen35-ggml/terminal.json", "sha256": "27f7b57552fd77e6afe48a4fe35bfec548230d7861fb06f68db76b6b5d7f4584", "bytes": 158069, "expected_sha256": "27f7b57552fd77e6afe48a4fe35bfec548230d7861fb06f68db76b6b5d7f4584", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "ollama-qwen35-llamacpp/terminal.json", "sha256": "9a0669f2a0dd0c52705a43ce2bdb91344f2b3f3b3ebd2e8dce37951c3eb40e3a", "bytes": 159292, "expected_sha256": "9a0669f2a0dd0c52705a43ce2bdb91344f2b3f3b3ebd2e8dce37951c3eb40e3a", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "ollama-qwen3vl/terminal.json", "sha256": "3e0f466426745c064f2e3604b1c76ab7685a23ee38ea8488b4f320eb5d86a7ff", "bytes": 178766, "expected_sha256": "3e0f466426745c064f2e3604b1c76ab7685a23ee38ea8488b4f320eb5d86a7ff", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "ollama-embedding/terminal.json", "sha256": "2901568a3f20812d8193046eb4cc9baa023a81659923aba715b017b797e0142b", "bytes": 88038, "expected_sha256": "2901568a3f20812d8193046eb4cc9baa023a81659923aba715b017b797e0142b", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "strata-iq2_xs/terminal.json", "sha256": "66c62851c0a2bfdf721131041cabeefdcee3d9b1809288954043df3983145d39", "bytes": 49016, "expected_sha256": "66c62851c0a2bfdf721131041cabeefdcee3d9b1809288954043df3983145d39", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "strata-iq3_xxs/terminal.json", "sha256": "72b7787f8a501d9b3da055390fbc683a1144a52fc9f10b6d6aaaf8420117c31c", "bytes": 132442, "expected_sha256": "72b7787f8a501d9b3da055390fbc683a1144a52fc9f10b6d6aaaf8420117c31c", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "strata-iq3_s/terminal.json", "sha256": "7bc6aeaa857413a04eb9ac56337a3c9453b4947fbb58123fd494d9e9fec1835f", "bytes": 140019, "expected_sha256": "7bc6aeaa857413a04eb9ac56337a3c9453b4947fbb58123fd494d9e9fec1835f", "recorded_digest_matches": true }, { "source_study": "original-v3", "artifact": "strata-ud-q4_k_xl/terminal.json", "sha256": "946ab21541a510acfa3fe4c0095db6e300bc5aecea93e5dbc394abc2a2884873", "bytes": 79406, "expected_sha256": "946ab21541a510acfa3fe4c0095db6e300bc5aecea93e5dbc394abc2a2884873", "recorded_digest_matches": true }, { "source_study": "iq2-fresh-tail4", "artifact": "strata-iq2_xs/terminal.json", "sha256": "ec85f7dfbc9c6efec1d85fb032890fcffb279d2f8dde88fe37b0fc78ecc0408b", "bytes": 45696, "expected_sha256": "ec85f7dfbc9c6efec1d85fb032890fcffb279d2f8dde88fe37b0fc78ecc0408b", "recorded_digest_matches": true }, { "source_study": "q4-tail9", "artifact": "strata-ud-q4_k_xl/terminal.json", "sha256": "b3b9210022f3409412702ed090f768b0fbede5e433bfeb6e1c2a56cc2710bdb2", "bytes": 81274, "expected_sha256": "b3b9210022f3409412702ed090f768b0fbede5e433bfeb6e1c2a56cc2710bdb2", "recorded_digest_matches": true } ], "limitations": [ "179 is canonical attempted-or-intent coverage, not 179 completed responses. The interrupted IQ2 pending call has persisted intent but no finalized response or verified dispatch.", "The interrupted IQ2 supplement has nine persisted completed response records and one unfinalized pending intent. Its lifecycle and historical release are unverified; all nine records are diagnostic only, never formal quality or speed samples.", "Capture-time idle/process observations do not prove historical unload, owned stop, watchdog join, resource qualification, or recovered baseline for the interrupted lifecycle.", "IQ2 tail4 copies only original requests[13:] (partial supplement requests[10:]). It never repeats the original failed cell, the interrupted nine records, or the pending call.", "Original IQ2 malformed-output failure and original Q4 deadline-stop failure remain failures. Original v3 execution and all-options successful execution remain false, even if attempted-or-intent coverage reaches 179.", "Formal aggregates use only separately completed lifecycle-qualified sealed segments. No candidate-level quality, cold-load, memory, or timing pooling across studies; no global intelligence ranking.", "Persisted all-attempt counters include noncanonical original Q4 interruption counters. They are not API-usage, confirmed-dispatch, token-cost or total-compute measurements; interrupted last dispatch may be unproven.", "Response records include failed/truncated records; recorded, delivered, deterministic-correct, lifecycle-qualified and human-accepted are distinct gates.", "QA grades are not imported here. Separate blind-review artifacts may exist elsewhere; diagnostic QA stays ungraded and human/product acceptance remains pending.", "Speed groups contain only the three frozen repetitions from one study/candidate/track/case; medians and ranges are descriptive, not p95, confidence intervals or a universal ranking.", "Cold means model residency-cold, not disk/OS-cache cold. Fixed order and warm prompt reuse can affect results. No cross-study load, lifecycle, memory or timing pooling.", "Memory is whole-system sampling around 1 Hz, not model-exclusive allocation or exact transient peaks.", "Context documents are 6K/24K characters, not tokens or proof of 32K-token context capacity.", "Workflow/context/code tests are synthetic bounded fixtures; no live Unreal analysis, profiling, edits or product acceptance is established.", "Vision fixtures are assisted PNG plus RenderSpec/host statistics with reused developer labels, not independent pure-vision holdout or rendering/profiling proof.", "Embedding retrieval uses eight developer-authored queries, not full Chroma/RAG acceptance. GGML/llamacpp labels identify artifact configurations, not distinct actual engines.", "Source bytes, protected hashes and installed asset size/mtime bindings are checked; model-weight bytes are not rehashed. This postprocessor does not check live machine state." ], "publication_exclusions": [ "Absolute local paths, process identifiers, private anonymous-review label mapping, unrestricted raw transcripts, prompts, generated source and full model-weight byte claims are omitted.", "Original studies and frozen artifacts are unchanged. Public draft contains derived measurements and scope metadata only." ], "cell_evidence": [ { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "code--higher_ring_buffer_minimal_fix--0", "case_id": "higher_ring_buffer_minimal_fix", "track": "code", "repeat": 0, "frozen_request_sha256": "e55bb62f3a2d23843031578be643c9aeb3ec27f0186701859b4bc5eb4ab0344f", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 76.09631819999777, "ttft_any_seconds": 1.4516297000227496, "ttft_reasoning_seconds": 1.4516297000227496, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 164.55131364254393, "api_prompt_tokens_per_second": 679.1597452934662, "sampled_peak_gpu_used_mib": 14847, "sampled_peak_system_ram_used_gib": 21.190189361572266, "sampled_min_available_phys_gib": 40.465362548828125, "sampled_min_available_commit_gib": 26.408004760742188, "resource_sample_count": 72, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "code--higher_gpu_interval_percentile--0", "case_id": "higher_gpu_interval_percentile", "track": "code", "repeat": 0, "frozen_request_sha256": "25931c2a96879416d96e11efca525702476bdfd0c6f06589b8c27ea48f36c166", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 75.33018190000439, "ttft_any_seconds": 0.22097779999603517, "ttft_reasoning_seconds": 0.22097779999603517, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 163.560142054539, "api_prompt_tokens_per_second": 8788.150294526946, "sampled_peak_gpu_used_mib": 14870, "sampled_peak_system_ram_used_gib": 21.18798065185547, "sampled_min_available_phys_gib": 40.46757125854492, "sampled_min_available_commit_gib": 26.316909790039062, "resource_sample_count": 71, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "code--higher_safe_proposal_validator--0", "case_id": "higher_safe_proposal_validator", "track": "code", "repeat": 0, "frozen_request_sha256": "0ed50be588a8e5eb98468d41d54ffe72523d711de8f1e94154186d8223bd38e7", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 74.60406810001587, "ttft_any_seconds": 0.24252730002626777, "ttft_reasoning_seconds": 0.24252730002626777, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 165.20665997952557, "api_prompt_tokens_per_second": 8551.49185227108, "sampled_peak_gpu_used_mib": 14802, "sampled_peak_system_ram_used_gib": 21.258956909179688, "sampled_min_available_phys_gib": 40.3965950012207, "sampled_min_available_commit_gib": 26.486553192138672, "resource_sample_count": 71, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "code--higher_bilingual_capture_status--0", "case_id": "higher_bilingual_capture_status", "track": "code", "repeat": 0, "frozen_request_sha256": "07bf92108bc84096521d33e402038267c87a5c11e679cc618e69406b90b64099", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 74.65054319999763, "ttft_any_seconds": 0.24601350000011735, "ttft_reasoning_seconds": 0.24601350000011735, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 165.10666531496332, "api_prompt_tokens_per_second": 7827.890483872166, "sampled_peak_gpu_used_mib": 14799, "sampled_peak_system_ram_used_gib": 21.52623748779297, "sampled_min_available_phys_gib": 40.12931442260742, "sampled_min_available_commit_gib": 26.203948974609375, "resource_sample_count": 71, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "qa--Q1-Chinese-conversation--0", "case_id": "Q1-Chinese-conversation", "track": "qa", "repeat": 0, "frozen_request_sha256": "261f425ce0982aa7fc4d4a43c7d9f2a797930cc5dc14c78f271e58c6304ad056", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 74.66623799997615, "ttft_any_seconds": 0.16065320000052452, "ttft_reasoning_seconds": 0.16065320000052452, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 164.88605914401754, "api_prompt_tokens_per_second": 5430.566330488751, "sampled_peak_gpu_used_mib": 14878, "sampled_peak_system_ram_used_gib": 21.90402603149414, "sampled_min_available_phys_gib": 39.75152587890625, "sampled_min_available_commit_gib": 25.784191131591797, "resource_sample_count": 71, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "qa--Q2-Unpaired-percentiles--0", "case_id": "Q2-Unpaired-percentiles", "track": "qa", "repeat": 0, "frozen_request_sha256": "11ccc77c8d32b8b01c2589b69ce045dadfd9b0baa6cddd4204e1038dcc049031", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 74.28841589999502, "ttft_any_seconds": 0.1493964999972377, "ttft_reasoning_seconds": 0.1493964999972377, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 165.69857239784335, "api_prompt_tokens_per_second": 5872.889733103328, "sampled_peak_gpu_used_mib": 14798, "sampled_peak_system_ram_used_gib": 22.158695220947266, "sampled_min_available_phys_gib": 39.496856689453125, "sampled_min_available_commit_gib": 25.548503875732422, "resource_sample_count": 70, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "qa--Q3-Overlapping-GPU-scopes--0", "case_id": "Q3-Overlapping-GPU-scopes", "track": "qa", "repeat": 0, "frozen_request_sha256": "af8967117a5e2a84ee4e0f6cf7cc04427b48ba99e5adae2308f5a5f451896a27", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 12.04581350000808, "ttft_any_seconds": 0.1608622000203468, "ttft_reasoning_seconds": 0.1608622000203468, "ttft_answer_seconds": 9.506439000018872, "api_decode_tokens_per_second": 170.9393955582973, "api_prompt_tokens_per_second": 6201.228936555208, "sampled_peak_gpu_used_mib": 14795, "sampled_peak_system_ram_used_gib": 22.425682067871094, "sampled_min_available_phys_gib": 39.2298698425293, "sampled_min_available_commit_gib": 25.306655883789062, "resource_sample_count": 12, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "qa--Q4-Untrusted-version-advice--0", "case_id": "Q4-Untrusted-version-advice", "track": "qa", "repeat": 0, "frozen_request_sha256": "2522f2369d86ccac53bc29da8480ba8aff63206fcc2b1d1f686f5ab0c76dcc96", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 74.87962629998219, "ttft_any_seconds": 0.10158269997918978, "ttft_reasoning_seconds": 0.10158269997918978, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 164.28339650627987, "api_prompt_tokens_per_second": 6141.1081462878665, "sampled_peak_gpu_used_mib": 14806, "sampled_peak_system_ram_used_gib": 22.755821228027344, "sampled_min_available_phys_gib": 38.89973068237305, "sampled_min_available_commit_gib": 24.86345672607422, "resource_sample_count": 71, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "context--context_lookup_6k_chars--0", "case_id": "context_lookup_6k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "f80c0b7d8622c8638053e494128851509f87c848b294cb4edb12e0bcab3e33ec", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.1400362999993376, "ttft_any_seconds": 0.38584269999410026, "ttft_reasoning_seconds": 0.38584269999410026, "ttft_answer_seconds": 0.8282952999870759, "api_decode_tokens_per_second": 169.2123873801143, "api_prompt_tokens_per_second": 10012.851315862994, "sampled_peak_gpu_used_mib": 14787, "sampled_peak_system_ram_used_gib": 22.841617584228516, "sampled_min_available_phys_gib": 38.813934326171875, "sampled_min_available_commit_gib": 24.8939208984375, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "context--context_lookup_24k_chars--0", "case_id": "context_lookup_24k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "9959b3b338c9df1228fbae7669d09a432c861b7be46ea57575d5dd062f50b6a8", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.7369869999820367, "ttft_any_seconds": 1.0403913999907672, "ttft_reasoning_seconds": 1.0403913999907672, "ttft_answer_seconds": 1.461492799979169, "api_decode_tokens_per_second": 164.3465483742923, "api_prompt_tokens_per_second": 10510.158535449707, "sampled_peak_gpu_used_mib": 14789, "sampled_peak_system_ram_used_gib": 22.940418243408203, "sampled_min_available_phys_gib": 38.71513366699219, "sampled_min_available_commit_gib": 24.80887222290039, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "workflow--workflow_no_pie_profile--0", "case_id": "workflow_no_pie_profile", "track": "workflow", "repeat": 0, "frozen_request_sha256": "236d8f10728fbced82a3d44cd36a49abbdaac56145416463160c7d07da18ac6f", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 5.581046799983596, "ttft_any_seconds": 0.18356430000858381, "ttft_reasoning_seconds": 0.18356430000858381, "ttft_answer_seconds": 5.081621899997117, "api_decode_tokens_per_second": 170.40231747151762, "api_prompt_tokens_per_second": 7083.534265659326, "sampled_peak_gpu_used_mib": 14787, "sampled_peak_system_ram_used_gib": 23.186199188232422, "sampled_min_available_phys_gib": 38.46935272216797, "sampled_min_available_commit_gib": 24.55249786376953, "resource_sample_count": 6, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "workflow--workflow_untrusted_excerpt--0", "case_id": "workflow_untrusted_excerpt", "track": "workflow", "repeat": 0, "frozen_request_sha256": "65662f0cc8d8f8ed1dbaf0c84000666dba1830634f669cce60c7f1b23ab470f1", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 6.5549167999997735, "ttft_any_seconds": 0.14334869998856448, "ttft_reasoning_seconds": 0.14334869998856448, "ttft_answer_seconds": 5.865306099993177, "api_decode_tokens_per_second": 165.83513376616253, "api_prompt_tokens_per_second": 7779.237249750883, "sampled_peak_gpu_used_mib": 14788, "sampled_peak_system_ram_used_gib": 23.233089447021484, "sampled_min_available_phys_gib": 38.422462463378906, "sampled_min_available_commit_gib": 24.530593872070312, "resource_sample_count": 7, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "workflow--workflow_read_only_capture--0", "case_id": "workflow_read_only_capture", "track": "workflow", "repeat": 0, "frozen_request_sha256": "050a4de4c2a4ea3a04b825c27fef0bc7b242865a99970204eef9906ae5ec3bb1", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 26.746131100022467, "ttft_any_seconds": 2.4639328000193927, "ttft_reasoning_seconds": 2.4639328000193927, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 168.54833018436824, "api_prompt_tokens_per_second": 293.25918045563526, "sampled_peak_gpu_used_mib": 14796, "sampled_peak_system_ram_used_gib": 23.344181060791016, "sampled_min_available_phys_gib": 38.311370849609375, "sampled_min_available_commit_gib": 24.401107788085938, "resource_sample_count": 26, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "workflow--workflow_destructive_approval--0", "case_id": "workflow_destructive_approval", "track": "workflow", "repeat": 0, "frozen_request_sha256": "0fa0c3bbacb67c85d69f680ec7dc01db76c91bf2d35b31828df3afb99388d81c", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 4.235048500006087, "ttft_any_seconds": 0.12709880000329576, "ttft_reasoning_seconds": 0.12709880000329576, "ttft_answer_seconds": 3.6585982000106014, "api_decode_tokens_per_second": 172.5043980559589, "api_prompt_tokens_per_second": 10866.284334440084, "sampled_peak_gpu_used_mib": 14787, "sampled_peak_system_ram_used_gib": 23.426437377929688, "sampled_min_available_phys_gib": 38.2291145324707, "sampled_min_available_commit_gib": 24.310985565185547, "resource_sample_count": 4, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "performance--speed_graphics_queue_512--1", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 1, "frozen_request_sha256": "ea4b482406b94d52dae81fbf98e0c674711396dab01b494a0d32bb5437dc8c6a", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 3.071733300021151, "ttft_any_seconds": 0.12060290001682006, "ttft_reasoning_seconds": 0.12060290001682006, "ttft_answer_seconds": 0.19186330001684837, "api_decode_tokens_per_second": 172.28476518124896, "api_prompt_tokens_per_second": 5207.622242593279, "sampled_peak_gpu_used_mib": 14798, "sampled_peak_system_ram_used_gib": 23.46326446533203, "sampled_min_available_phys_gib": 38.19228744506836, "sampled_min_available_commit_gib": 24.262500762939453, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "performance--speed_graphics_queue_512--2", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 2, "frozen_request_sha256": "ea4b482406b94d52dae81fbf98e0c674711396dab01b494a0d32bb5437dc8c6a", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 3.023247699980857, "ttft_any_seconds": 0.07289459998719394, "ttft_reasoning_seconds": 0.07289459998719394, "ttft_answer_seconds": 0.1416540999780409, "api_decode_tokens_per_second": 172.42511546926067, "api_prompt_tokens_per_second": 28069.806560134566, "sampled_peak_gpu_used_mib": 14802, "sampled_peak_system_ram_used_gib": 23.465232849121094, "sampled_min_available_phys_gib": 38.1903190612793, "sampled_min_available_commit_gib": 24.263469696044922, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "request_id": "performance--speed_graphics_queue_512--3", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 3, "frozen_request_sha256": "ea4b482406b94d52dae81fbf98e0c674711396dab01b494a0d32bb5437dc8c6a", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.998282199987443, "ttft_any_seconds": 0.05758589998004027, "ttft_reasoning_seconds": 0.05758589998004027, "ttft_answer_seconds": 0.12600279998150654, "api_decode_tokens_per_second": 172.89639622451003, "api_prompt_tokens_per_second": 21774.588158538576, "sampled_peak_gpu_used_mib": 14789, "sampled_peak_system_ram_used_gib": 23.47713851928711, "sampled_min_available_phys_gib": 38.17841339111328, "sampled_min_available_commit_gib": 24.27869415283203, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "code--higher_ring_buffer_minimal_fix--0", "case_id": "higher_ring_buffer_minimal_fix", "track": "code", "repeat": 0, "frozen_request_sha256": "943020349a3f2881d4cec24c6e7e884ee721e9de23844293c9e21cc9322b3bc5", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 74.62186620000284, "ttft_any_seconds": 0.3728017999965232, "ttft_reasoning_seconds": 0.3728017999965232, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 165.4455386556271, "api_prompt_tokens_per_second": 2925.365903649254, "sampled_peak_gpu_used_mib": 14798, "sampled_peak_system_ram_used_gib": 20.597087860107422, "sampled_min_available_phys_gib": 41.05846405029297, "sampled_min_available_commit_gib": 27.242050170898438, "resource_sample_count": 71, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "code--higher_gpu_interval_percentile--0", "case_id": "higher_gpu_interval_percentile", "track": "code", "repeat": 0, "frozen_request_sha256": "bda35ca520dd1e49213b2a306186d38cef8a8969049f33aaf943e578b6c888e3", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 73.06745610001963, "ttft_any_seconds": 0.2062836000113748, "ttft_reasoning_seconds": 0.2062836000113748, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 168.60693623249446, "api_prompt_tokens_per_second": 8795.02442002442, "sampled_peak_gpu_used_mib": 14875, "sampled_peak_system_ram_used_gib": 20.945579528808594, "sampled_min_available_phys_gib": 40.7099723815918, "sampled_min_available_commit_gib": 26.868244171142578, "resource_sample_count": 69, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "code--higher_safe_proposal_validator--0", "case_id": "higher_safe_proposal_validator", "track": "code", "repeat": 0, "frozen_request_sha256": "b9800febab1e92afb29d64240a6cc64f3d3d1d84969f5b90a2540937f8d3aae3", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 72.17519229999743, "ttft_any_seconds": 0.24071589999948628, "ttft_reasoning_seconds": 0.24071589999948628, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 170.77990323028075, "api_prompt_tokens_per_second": 8843.556837161497, "sampled_peak_gpu_used_mib": 14812, "sampled_peak_system_ram_used_gib": 21.23507308959961, "sampled_min_available_phys_gib": 40.42047882080078, "sampled_min_available_commit_gib": 26.59100341796875, "resource_sample_count": 68, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "code--higher_bilingual_capture_status--0", "case_id": "higher_bilingual_capture_status", "track": "code", "repeat": 0, "frozen_request_sha256": "63c5b8c9b5b689ee837ad12a5c45a8905cc137b401c92a54b473f0a8d76341cf", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 72.14386499999091, "ttft_any_seconds": 0.25549640000099316, "ttft_reasoning_seconds": 0.25549640000099316, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 170.889701461755, "api_prompt_tokens_per_second": 8313.004760438624, "sampled_peak_gpu_used_mib": 14819, "sampled_peak_system_ram_used_gib": 21.542709350585938, "sampled_min_available_phys_gib": 40.11284255981445, "sampled_min_available_commit_gib": 26.24233627319336, "resource_sample_count": 68, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "qa--Q1-Chinese-conversation--0", "case_id": "Q1-Chinese-conversation", "track": "qa", "repeat": 0, "frozen_request_sha256": "bcd125f047bd9e1fe22db49daa69f41997d2bb8c735199d952b78ca6e617a4a9", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 71.66968130000168, "ttft_any_seconds": 0.18984830001136288, "ttft_reasoning_seconds": 0.18984830001136288, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 171.86626695264982, "api_prompt_tokens_per_second": 5292.268512991923, "sampled_peak_gpu_used_mib": 14819, "sampled_peak_system_ram_used_gib": 21.65152359008789, "sampled_min_available_phys_gib": 40.0040283203125, "sampled_min_available_commit_gib": 25.96611785888672, "resource_sample_count": 68, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "qa--Q2-Unpaired-percentiles--0", "case_id": "Q2-Unpaired-percentiles", "track": "qa", "repeat": 0, "frozen_request_sha256": "f8417afa0ac3b14afc3818ae16cf70d8cd84ecaf7e7af58ca9f1b7f82d2e1c03", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 71.70253770001, "ttft_any_seconds": 0.1638643000042066, "ttft_reasoning_seconds": 0.1638643000042066, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 171.723839166482, "api_prompt_tokens_per_second": 5826.2508172133985, "sampled_peak_gpu_used_mib": 14819, "sampled_peak_system_ram_used_gib": 21.94662857055664, "sampled_min_available_phys_gib": 39.70892333984375, "sampled_min_available_commit_gib": 25.71877670288086, "resource_sample_count": 68, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "qa--Q3-Overlapping-GPU-scopes--0", "case_id": "Q3-Overlapping-GPU-scopes", "track": "qa", "repeat": 0, "frozen_request_sha256": "fc8d6bba7f2a362c65cdcd8a7db393ff9efe58c510c8ba7c9bc4a755af900f7e", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 11.82708650000859, "ttft_any_seconds": 0.17282069998327643, "ttft_reasoning_seconds": 0.17282069998327643, "ttft_answer_seconds": 9.368044100003317, "api_decode_tokens_per_second": 174.31329555603028, "api_prompt_tokens_per_second": 6307.515337423313, "sampled_peak_gpu_used_mib": 14903, "sampled_peak_system_ram_used_gib": 22.42003631591797, "sampled_min_available_phys_gib": 39.23551559448242, "sampled_min_available_commit_gib": 25.08179473876953, "resource_sample_count": 12, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "qa--Q4-Untrusted-version-advice--0", "case_id": "Q4-Untrusted-version-advice", "track": "qa", "repeat": 0, "frozen_request_sha256": "7fd69fb2e19bdec6b042fc816a9f3374d6daa2049ef7f57f904f8fd9775a57c3", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 73.77385369999683, "ttft_any_seconds": 0.10302019998198375, "ttft_reasoning_seconds": 0.10302019998198375, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 166.7541442876331, "api_prompt_tokens_per_second": 6291.457286432161, "sampled_peak_gpu_used_mib": 14818, "sampled_peak_system_ram_used_gib": 22.456336975097656, "sampled_min_available_phys_gib": 39.199214935302734, "sampled_min_available_commit_gib": 25.122196197509766, "resource_sample_count": 70, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "context--context_lookup_6k_chars--0", "case_id": "context_lookup_6k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "9bfb9aef45622fe5bdf8f60abfd786a019bf22b4b849631a8e393a163944a03b", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.2215485000051558, "ttft_any_seconds": 0.4195451999839861, "ttft_reasoning_seconds": 0.4195451999839861, "ttft_answer_seconds": 0.8971531999995932, "api_decode_tokens_per_second": 158.9096131824622, "api_prompt_tokens_per_second": 9153.473927242141, "sampled_peak_gpu_used_mib": 14782, "sampled_peak_system_ram_used_gib": 22.670021057128906, "sampled_min_available_phys_gib": 38.985530853271484, "sampled_min_available_commit_gib": 24.923004150390625, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "context--context_lookup_24k_chars--0", "case_id": "context_lookup_24k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "31df7ae3219b8c47564c9679891494ef4701184690678c1e4b7d86cd2cf55a1a", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.8460401000047568, "ttft_any_seconds": 1.119779100001324, "ttft_reasoning_seconds": 1.119779100001324, "ttft_answer_seconds": 1.5685476000071503, "api_decode_tokens_per_second": 157.80274389550783, "api_prompt_tokens_per_second": 9533.416970757407, "sampled_peak_gpu_used_mib": 14782, "sampled_peak_system_ram_used_gib": 22.754703521728516, "sampled_min_available_phys_gib": 38.900848388671875, "sampled_min_available_commit_gib": 24.836944580078125, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "workflow--workflow_no_pie_profile--0", "case_id": "workflow_no_pie_profile", "track": "workflow", "repeat": 0, "frozen_request_sha256": "5733738a147b84db78b8d12d849a3920d51f4aa0485469d8e4c2fb422effc6dd", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 5.877615200006403, "ttft_any_seconds": 0.20286370001849718, "ttft_reasoning_seconds": 0.20286370001849718, "ttft_answer_seconds": 5.364163899997948, "api_decode_tokens_per_second": 162.10922543755441, "api_prompt_tokens_per_second": 7042.778754461669, "sampled_peak_gpu_used_mib": 14797, "sampled_peak_system_ram_used_gib": 22.991012573242188, "sampled_min_available_phys_gib": 38.6645393371582, "sampled_min_available_commit_gib": 24.55902099609375, "resource_sample_count": 6, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "workflow--workflow_untrusted_excerpt--0", "case_id": "workflow_untrusted_excerpt", "track": "workflow", "repeat": 0, "frozen_request_sha256": "4bc53ab55c65e83a4f8b8db318f9040bf74bc151ff29a655545550b7db2a1a5a", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 6.160213700000895, "ttft_any_seconds": 0.15603959999862127, "ttft_reasoning_seconds": 0.15603959999862127, "ttft_answer_seconds": 5.493975799996406, "api_decode_tokens_per_second": 177.11889475153723, "api_prompt_tokens_per_second": 8548.2847437381, "sampled_peak_gpu_used_mib": 14789, "sampled_peak_system_ram_used_gib": 23.028884887695312, "sampled_min_available_phys_gib": 38.62666702270508, "sampled_min_available_commit_gib": 24.552692413330078, "resource_sample_count": 6, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "workflow--workflow_read_only_capture--0", "case_id": "workflow_read_only_capture", "track": "workflow", "repeat": 0, "frozen_request_sha256": "ab352021fbed914888de6fa2276413d12a02c148cd352213d01bbc54ef734c7d", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 23.51732740001171, "ttft_any_seconds": 0.13569480000296608, "ttft_reasoning_seconds": 0.13569480000296608, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 175.03589069059908, "api_prompt_tokens_per_second": 8393.802236750724, "sampled_peak_gpu_used_mib": 14791, "sampled_peak_system_ram_used_gib": 23.0804443359375, "sampled_min_available_phys_gib": 38.57510757446289, "sampled_min_available_commit_gib": 24.477794647216797, "resource_sample_count": 23, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "workflow--workflow_destructive_approval--0", "case_id": "workflow_destructive_approval", "track": "workflow", "repeat": 0, "frozen_request_sha256": "7121244aeb1c1e50eb7e0c6aa998b8a56fc26c6a3ee45cac5c40eefefd55403a", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 4.141411899996456, "ttft_any_seconds": 0.14609210001071915, "ttft_reasoning_seconds": 0.14609210001071915, "ttft_answer_seconds": 3.5828130999871064, "api_decode_tokens_per_second": 177.33737888380068, "api_prompt_tokens_per_second": 11170.994610485057, "sampled_peak_gpu_used_mib": 14791, "sampled_peak_system_ram_used_gib": 23.168060302734375, "sampled_min_available_phys_gib": 38.487491607666016, "sampled_min_available_commit_gib": 24.3907470703125, "resource_sample_count": 4, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "performance--speed_graphics_queue_512--1", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 1, "frozen_request_sha256": "8e3a88d61cdb5f7ca481a3b82223880429ef28332880f69a9d0f8bcc62be93bd", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.9347764999838546, "ttft_any_seconds": 0.09583599999314174, "ttft_reasoning_seconds": 0.09583599999314174, "ttft_answer_seconds": 0.16333330000634305, "api_decode_tokens_per_second": 179.09249077778608, "api_prompt_tokens_per_second": 5419.889166311431, "sampled_peak_gpu_used_mib": 14791, "sampled_peak_system_ram_used_gib": 23.195819854736328, "sampled_min_available_phys_gib": 38.45973205566406, "sampled_min_available_commit_gib": 24.359291076660156, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "performance--speed_graphics_queue_512--2", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 2, "frozen_request_sha256": "8e3a88d61cdb5f7ca481a3b82223880429ef28332880f69a9d0f8bcc62be93bd", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.88999560000957, "ttft_any_seconds": 0.05173020000802353, "ttft_reasoning_seconds": 0.05173020000802353, "ttft_answer_seconds": 0.11897830001544207, "api_decode_tokens_per_second": 179.24440780203295, "api_prompt_tokens_per_second": 27946.40988067825, "sampled_peak_gpu_used_mib": 14791, "sampled_peak_system_ram_used_gib": 23.218303680419922, "sampled_min_available_phys_gib": 38.43724822998047, "sampled_min_available_commit_gib": 24.332401275634766, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "request_id": "performance--speed_graphics_queue_512--3", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 3, "frozen_request_sha256": "8e3a88d61cdb5f7ca481a3b82223880429ef28332880f69a9d0f8bcc62be93bd", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.9008435000141617, "ttft_any_seconds": 0.05597010001656599, "ttft_reasoning_seconds": 0.05597010001656599, "ttft_answer_seconds": 0.12228010001126677, "api_decode_tokens_per_second": 178.7301987361122, "api_prompt_tokens_per_second": 22225.92191792225, "sampled_peak_gpu_used_mib": 14790, "sampled_peak_system_ram_used_gib": 23.245346069335938, "sampled_min_available_phys_gib": 38.41020584106445, "sampled_min_available_commit_gib": 24.305580139160156, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "code--higher_ring_buffer_minimal_fix--0", "case_id": "higher_ring_buffer_minimal_fix", "track": "code", "repeat": 0, "frozen_request_sha256": "612faa0d14be99cd0c04afc6878d263265592b567f33c29604d1c7da8784bcbd", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 8.439725400006864, "ttft_any_seconds": 0.2432437999814283, "ttft_reasoning_seconds": 0.2432437999814283, "ttft_answer_seconds": 6.768799699988449, "api_decode_tokens_per_second": 130.1785947836815, "api_prompt_tokens_per_second": 4404.156576890261, "sampled_peak_gpu_used_mib": 9117, "sampled_peak_system_ram_used_gib": 20.39282989501953, "sampled_min_available_phys_gib": 41.26272201538086, "sampled_min_available_commit_gib": 31.63784408569336, "resource_sample_count": 8, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "code--higher_gpu_interval_percentile--0", "case_id": "higher_gpu_interval_percentile", "track": "code", "repeat": 0, "frozen_request_sha256": "fad718ce978678aa33c93a6420a4bb58e91f4b4dc7a236a35468d1d03beec66c", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 89.11300149999443, "ttft_any_seconds": 0.188392499985639, "ttft_reasoning_seconds": 0.188392499985639, "ttft_answer_seconds": 79.31756719999248, "api_decode_tokens_per_second": 127.6602994517928, "api_prompt_tokens_per_second": 5942.021103013162, "sampled_peak_gpu_used_mib": 9124, "sampled_peak_system_ram_used_gib": 20.36505889892578, "sampled_min_available_phys_gib": 41.29049301147461, "sampled_min_available_commit_gib": 31.64128875732422, "resource_sample_count": 84, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "code--higher_safe_proposal_validator--0", "case_id": "higher_safe_proposal_validator", "track": "code", "repeat": 0, "frozen_request_sha256": "c42dfbc59b21651ca93034fec8ca8b66ef24c4f3097a9e1b97d3f706acbb5e12", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 97.8131555999862, "ttft_any_seconds": 0.32556200001272373, "ttft_reasoning_seconds": 0.32556200001272373, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 126.04809405649694, "api_prompt_tokens_per_second": 5727.249420744833, "sampled_peak_gpu_used_mib": 9125, "sampled_peak_system_ram_used_gib": 20.84805679321289, "sampled_min_available_phys_gib": 40.8074951171875, "sampled_min_available_commit_gib": 31.131954193115234, "resource_sample_count": 92, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "code--higher_bilingual_capture_status--0", "case_id": "higher_bilingual_capture_status", "track": "code", "repeat": 0, "frozen_request_sha256": "ee676f253e4c6f927f41e55312cadf9c58196320a46263b389362e66ecea0936", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 97.08686999999918, "ttft_any_seconds": 0.29483020000043325, "ttft_reasoning_seconds": 0.29483020000043325, "ttft_answer_seconds": 89.51853450000635, "api_decode_tokens_per_second": 126.95396282331728, "api_prompt_tokens_per_second": 6324.984933115459, "sampled_peak_gpu_used_mib": 9132, "sampled_peak_system_ram_used_gib": 21.40973663330078, "sampled_min_available_phys_gib": 40.24581527709961, "sampled_min_available_commit_gib": 30.569828033447266, "resource_sample_count": 92, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "qa--Q1-Chinese-conversation--0", "case_id": "Q1-Chinese-conversation", "track": "qa", "repeat": 0, "frozen_request_sha256": "774a66dde1ac65e5bb483e81fd3864ba93dab7edca807fe213f6b4fd3e596f99", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 16.419944299996132, "ttft_any_seconds": 0.20989220001501963, "ttft_reasoning_seconds": 0.20989220001501963, "ttft_answer_seconds": 15.09848700001021, "api_decode_tokens_per_second": 129.9194178820445, "api_prompt_tokens_per_second": 3308.6040760637843, "sampled_peak_gpu_used_mib": 9123, "sampled_peak_system_ram_used_gib": 21.844757080078125, "sampled_min_available_phys_gib": 39.810794830322266, "sampled_min_available_commit_gib": 30.13103485107422, "resource_sample_count": 16, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "qa--Q2-Unpaired-percentiles--0", "case_id": "Q2-Unpaired-percentiles", "track": "qa", "repeat": 0, "frozen_request_sha256": "1418d17051d7373654ddda9b0586d2d2be0223ea1ecb2cb05abd02802ebd8b3e", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 14.972761999990325, "ttft_any_seconds": 0.10547159999259748, "ttft_reasoning_seconds": 0.10547159999259748, "ttft_answer_seconds": 11.11710000000312, "api_decode_tokens_per_second": 130.4204266234434, "api_prompt_tokens_per_second": 3994.3347284329793, "sampled_peak_gpu_used_mib": 9124, "sampled_peak_system_ram_used_gib": 22.01634979248047, "sampled_min_available_phys_gib": 39.63920211791992, "sampled_min_available_commit_gib": 29.954959869384766, "resource_sample_count": 15, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "qa--Q3-Overlapping-GPU-scopes--0", "case_id": "Q3-Overlapping-GPU-scopes", "track": "qa", "repeat": 0, "frozen_request_sha256": "2ed39f4bba51b3bd5abb71ba685411b968a85ea7123cb7d6d35db445f9a95f40", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 24.378757099999348, "ttft_any_seconds": 0.13270180000108667, "ttft_reasoning_seconds": 0.13270180000108667, "ttft_answer_seconds": 22.050559800001793, "api_decode_tokens_per_second": 129.58786727612284, "api_prompt_tokens_per_second": 4056.190163738853, "sampled_peak_gpu_used_mib": 9124, "sampled_peak_system_ram_used_gib": 22.191688537597656, "sampled_min_available_phys_gib": 39.463863372802734, "sampled_min_available_commit_gib": 29.772384643554688, "resource_sample_count": 23, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "qa--Q4-Untrusted-version-advice--0", "case_id": "Q4-Untrusted-version-advice", "track": "qa", "repeat": 0, "frozen_request_sha256": "d33f2e028858628f52bfe8091c4c0e7d2c8fe02222e5bd7bacaeae2910633b21", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 16.966344299988123, "ttft_any_seconds": 0.13727350000408478, "ttft_reasoning_seconds": 0.13727350000408478, "ttft_answer_seconds": 11.598757999978261, "api_decode_tokens_per_second": 130.19108735940804, "api_prompt_tokens_per_second": 4012.051766395554, "sampled_peak_gpu_used_mib": 9121, "sampled_peak_system_ram_used_gib": 22.381790161132812, "sampled_min_available_phys_gib": 39.27376174926758, "sampled_min_available_commit_gib": 29.62487030029297, "resource_sample_count": 16, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "context--context_lookup_6k_chars--0", "case_id": "context_lookup_6k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "400f1870c1592afb2236f15c416547be0b3547f7b1ab92b50b16c1abc3928ad9", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.1225361000106204, "ttft_any_seconds": 0.5983493000094313, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.5983493000094313, "api_decode_tokens_per_second": 131.62262077806136, "api_prompt_tokens_per_second": 6457.285481980108, "sampled_peak_gpu_used_mib": 9108, "sampled_peak_system_ram_used_gib": 22.595504760742188, "sampled_min_available_phys_gib": 39.0600471496582, "sampled_min_available_commit_gib": 29.419017791748047, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "context--context_lookup_24k_chars--0", "case_id": "context_lookup_24k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "354b766b8b06dcbc4eaf9df2c0eb87a692720ad3ebe6163dbb99e50ddc493527", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.578098099998897, "ttft_any_seconds": 2.0663832999998704, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 2.0663832999998704, "api_decode_tokens_per_second": 125.12390198888352, "api_prompt_tokens_per_second": 6574.476010916915, "sampled_peak_gpu_used_mib": 9106, "sampled_peak_system_ram_used_gib": 22.847579956054688, "sampled_min_available_phys_gib": 38.8079719543457, "sampled_min_available_commit_gib": 29.177913665771484, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "workflow--workflow_no_pie_profile--0", "case_id": "workflow_no_pie_profile", "track": "workflow", "repeat": 0, "frozen_request_sha256": "ad589d4918677ca2d4e6ff278e583afadc3bca7fa5632dafd7b2609e89b490ba", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 12.641738600010285, "ttft_any_seconds": 0.2621790000121109, "ttft_reasoning_seconds": 0.2621790000121109, "ttft_answer_seconds": 11.466203799995128, "api_decode_tokens_per_second": 130.3682077981997, "api_prompt_tokens_per_second": 5322.351257732678, "sampled_peak_gpu_used_mib": 9106, "sampled_peak_system_ram_used_gib": 23.388874053955078, "sampled_min_available_phys_gib": 38.26667785644531, "sampled_min_available_commit_gib": 28.660415649414062, "resource_sample_count": 12, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "workflow--workflow_untrusted_excerpt--0", "case_id": "workflow_untrusted_excerpt", "track": "workflow", "repeat": 0, "frozen_request_sha256": "bd901b326d9add78ab05988381bd5c88a23fc75570179d155f215f79b1ba7b92", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 23.29895990001387, "ttft_any_seconds": 0.17369510000571609, "ttft_reasoning_seconds": 0.17369510000571609, "ttft_answer_seconds": 21.732912300009048, "api_decode_tokens_per_second": 128.60691253938586, "api_prompt_tokens_per_second": 5694.160133727843, "sampled_peak_gpu_used_mib": 9117, "sampled_peak_system_ram_used_gib": 23.643043518066406, "sampled_min_available_phys_gib": 38.012508392333984, "sampled_min_available_commit_gib": 28.39855194091797, "resource_sample_count": 22, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "workflow--workflow_read_only_capture--0", "case_id": "workflow_read_only_capture", "track": "workflow", "repeat": 0, "frozen_request_sha256": "c29c5f99d4c89f4fe2bd70f179d20eab1fff5610e6291a3d08fc1a30f99711fc", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 28.594883400015533, "ttft_any_seconds": 0.19037250001565553, "ttft_reasoning_seconds": 0.19037250001565553, "ttft_answer_seconds": 26.85275449999608, "api_decode_tokens_per_second": 124.91227887349262, "api_prompt_tokens_per_second": 5612.634796637514, "sampled_peak_gpu_used_mib": 9212, "sampled_peak_system_ram_used_gib": 24.030811309814453, "sampled_min_available_phys_gib": 37.62474060058594, "sampled_min_available_commit_gib": 27.894546508789062, "resource_sample_count": 27, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "workflow--workflow_destructive_approval--0", "case_id": "workflow_destructive_approval", "track": "workflow", "repeat": 0, "frozen_request_sha256": "765afd4d912f6b7464d7393943707e72f7c3f5937ebe43fb795dbdf545c7d8c0", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 26.40692360000685, "ttft_any_seconds": 0.19205690000671893, "ttft_reasoning_seconds": 0.19205690000671893, "ttft_answer_seconds": 25.20529880002141, "api_decode_tokens_per_second": 127.75383617227992, "api_prompt_tokens_per_second": 5627.266908470237, "sampled_peak_gpu_used_mib": 9207, "sampled_peak_system_ram_used_gib": 25.01544189453125, "sampled_min_available_phys_gib": 36.64011001586914, "sampled_min_available_commit_gib": 26.902080535888672, "resource_sample_count": 25, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "vision--vision_oblique_bleached_wood--0", "case_id": "vision_oblique_bleached_wood", "track": "vision", "repeat": 0, "frozen_request_sha256": "75b3cfb15b9cbb3269c50180d43e6767f3ad645461f809c5b824aa743f227b10", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 6.067666799994186, "ttft_any_seconds": 5.442273299995577, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 5.442273299995577, "api_decode_tokens_per_second": 124.81078365170751, "api_prompt_tokens_per_second": 820.6682697392266, "sampled_peak_gpu_used_mib": 9391, "sampled_peak_system_ram_used_gib": 25.74044418334961, "sampled_min_available_phys_gib": 35.91510772705078, "sampled_min_available_commit_gib": 26.113697052001953, "resource_sample_count": 6, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "vision--vision_missing_wood_material--0", "case_id": "vision_missing_wood_material", "track": "vision", "repeat": 0, "frozen_request_sha256": "2d7d46b163700b39d0018c5e62e86523e6b275fcdeea7295edf7c68dff5e39df", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.8771076999837533, "ttft_any_seconds": 1.423017399996752, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.423017399996752, "api_decode_tokens_per_second": 126.58402014337017, "api_prompt_tokens_per_second": 3390.6987329285535, "sampled_peak_gpu_used_mib": 9391, "sampled_peak_system_ram_used_gib": 26.725059509277344, "sampled_min_available_phys_gib": 34.93049240112305, "sampled_min_available_commit_gib": 24.999282836914062, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "vision--vision_catalog_material_smoke--0", "case_id": "vision_catalog_material_smoke", "track": "vision", "repeat": 0, "frozen_request_sha256": "5dfdffbe30417b7e1afc372f358b4adda70d33965a50aaa4379c79cd8286e079", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.0111548000131734, "ttft_any_seconds": 1.4400070999981835, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.4400070999981835, "api_decode_tokens_per_second": 126.11555337578056, "api_prompt_tokens_per_second": 3404.795794778527, "sampled_peak_gpu_used_mib": 9391, "sampled_peak_system_ram_used_gib": 26.77178955078125, "sampled_min_available_phys_gib": 34.88376235961914, "sampled_min_available_commit_gib": 24.913543701171875, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "vision--vision_weathered_wood_fixed_exposure--0", "case_id": "vision_weathered_wood_fixed_exposure", "track": "vision", "repeat": 0, "frozen_request_sha256": "0dcf797e07c75e92270e5b03c25323be0004ca5d6b755627e517054f5d4647e7", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.8991194000118412, "ttft_any_seconds": 1.4140644000144675, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.4140644000144675, "api_decode_tokens_per_second": 127.91734063975176, "api_prompt_tokens_per_second": 3441.9937358381612, "sampled_peak_gpu_used_mib": 9391, "sampled_peak_system_ram_used_gib": 27.023040771484375, "sampled_min_available_phys_gib": 34.632511138916016, "sampled_min_available_commit_gib": 24.627544403076172, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "performance--speed_graphics_queue_512--1", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 1, "frozen_request_sha256": "3f7602fb2299f24e037beeb0f4ebc88c4862533e6d658b2ce09392ae64b2b153", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 4.130043699988164, "ttft_any_seconds": 0.13830089999828488, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.13830089999828488, "api_decode_tokens_per_second": 128.30273029212427, "api_prompt_tokens_per_second": 3399.778693651074, "sampled_peak_gpu_used_mib": 9391, "sampled_peak_system_ram_used_gib": 27.264530181884766, "sampled_min_available_phys_gib": 34.391021728515625, "sampled_min_available_commit_gib": 24.327613830566406, "resource_sample_count": 4, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "performance--speed_graphics_queue_512--2", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 2, "frozen_request_sha256": "3f7602fb2299f24e037beeb0f4ebc88c4862533e6d658b2ce09392ae64b2b153", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 4.126083200011635, "ttft_any_seconds": 0.06527240001014434, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.06527240001014434, "api_decode_tokens_per_second": 126.08564170203796, "api_prompt_tokens_per_second": 7332.595462091865, "sampled_peak_gpu_used_mib": 9403, "sampled_peak_system_ram_used_gib": 27.357128143310547, "sampled_min_available_phys_gib": 34.298423767089844, "sampled_min_available_commit_gib": 24.24774169921875, "resource_sample_count": 4, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "performance--speed_graphics_queue_512--3", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 3, "frozen_request_sha256": "3f7602fb2299f24e037beeb0f4ebc88c4862533e6d658b2ce09392ae64b2b153", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 4.0917132000031415, "ttft_any_seconds": 0.06464139997842722, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.06464139997842722, "api_decode_tokens_per_second": 127.14523437560142, "api_prompt_tokens_per_second": 7634.137558516384, "sampled_peak_gpu_used_mib": 9414, "sampled_peak_system_ram_used_gib": 27.36655044555664, "sampled_min_available_phys_gib": 34.28900146484375, "sampled_min_available_commit_gib": 24.230384826660156, "resource_sample_count": 4, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "vision-performance--vision_oblique_bleached_wood--1", "case_id": "vision_oblique_bleached_wood", "track": "vision-performance", "repeat": 1, "frozen_request_sha256": "75b3cfb15b9cbb3269c50180d43e6767f3ad645461f809c5b824aa743f227b10", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 0.7759215999976732, "ttft_any_seconds": 0.15225359998294152, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.15225359998294152, "api_decode_tokens_per_second": 125.09321850417378, "api_prompt_tokens_per_second": 144389.6301691885, "sampled_peak_gpu_used_mib": 9506, "sampled_peak_system_ram_used_gib": 27.70001983642578, "sampled_min_available_phys_gib": 33.95553207397461, "sampled_min_available_commit_gib": 23.68582534790039, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "vision-performance--vision_oblique_bleached_wood--2", "case_id": "vision_oblique_bleached_wood", "track": "vision-performance", "repeat": 2, "frozen_request_sha256": "75b3cfb15b9cbb3269c50180d43e6767f3ad645461f809c5b824aa743f227b10", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 0.7387351999932434, "ttft_any_seconds": 0.12193330001900904, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.12193330001900904, "api_decode_tokens_per_second": 126.46611518793189, "api_prompt_tokens_per_second": 150967.56326375186, "sampled_peak_gpu_used_mib": 9512, "sampled_peak_system_ram_used_gib": 27.506149291992188, "sampled_min_available_phys_gib": 34.1494026184082, "sampled_min_available_commit_gib": 23.89849853515625, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "request_id": "vision-performance--vision_oblique_bleached_wood--3", "case_id": "vision_oblique_bleached_wood", "track": "vision-performance", "repeat": 3, "frozen_request_sha256": "75b3cfb15b9cbb3269c50180d43e6767f3ad645461f809c5b824aa743f227b10", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 0.7119180999870878, "ttft_any_seconds": 0.10027649998664856, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.10027649998664856, "api_decode_tokens_per_second": 127.54664037243619, "api_prompt_tokens_per_second": 158311.24986386902, "sampled_peak_gpu_used_mib": 9512, "sampled_peak_system_ram_used_gib": 27.480133056640625, "sampled_min_available_phys_gib": 34.175418853759766, "sampled_min_available_commit_gib": 23.932147979736328, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "code--higher_ring_buffer_minimal_fix--0", "case_id": "higher_ring_buffer_minimal_fix", "track": "code", "repeat": 0, "frozen_request_sha256": "c1f015a2e6ef268e091c0d795e7fbe07f4a769250ce5606b5d67b67c257958a9", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 8.783852099993965, "ttft_any_seconds": 0.230253099987749, "ttft_reasoning_seconds": 0.230253099987749, "ttft_answer_seconds": 6.975197199994, "api_decode_tokens_per_second": 124.74934676121116, "api_prompt_tokens_per_second": 4685.917584522304, "sampled_peak_gpu_used_mib": 9266, "sampled_peak_system_ram_used_gib": 22.943161010742188, "sampled_min_available_phys_gib": 38.7123908996582, "sampled_min_available_commit_gib": 28.645645141601562, "resource_sample_count": 9, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "code--higher_gpu_interval_percentile--0", "case_id": "higher_gpu_interval_percentile", "track": "code", "repeat": 0, "frozen_request_sha256": "3907b3584d5b3b3f06a06d1b24331caa28d5d32f5f4a1a60703798cb3816ae8d", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 89.74120730001596, "ttft_any_seconds": 0.21701680001569912, "ttft_reasoning_seconds": 0.21701680001569912, "ttft_answer_seconds": 79.7897784000088, "api_decode_tokens_per_second": 126.80497561188008, "api_prompt_tokens_per_second": 5774.847039234602, "sampled_peak_gpu_used_mib": 9297, "sampled_peak_system_ram_used_gib": 23.050182342529297, "sampled_min_available_phys_gib": 38.605369567871094, "sampled_min_available_commit_gib": 28.500343322753906, "resource_sample_count": 85, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "code--higher_safe_proposal_validator--0", "case_id": "higher_safe_proposal_validator", "track": "code", "repeat": 0, "frozen_request_sha256": "404b7e751cdaab826ddc760686fba2602986e77b90aef6246e725cc814fb77b3", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 99.70993169999565, "ttft_any_seconds": 0.3233976000046823, "ttft_reasoning_seconds": 0.3233976000046823, "ttft_answer_seconds": null, "api_decode_tokens_per_second": 123.64080390429646, "api_prompt_tokens_per_second": 5777.610782270933, "sampled_peak_gpu_used_mib": 9269, "sampled_peak_system_ram_used_gib": 23.54636001586914, "sampled_min_available_phys_gib": 38.10919189453125, "sampled_min_available_commit_gib": 28.07135009765625, "resource_sample_count": 94, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "code--higher_bilingual_capture_status--0", "case_id": "higher_bilingual_capture_status", "track": "code", "repeat": 0, "frozen_request_sha256": "7ae3bffea1b969835bf6711bbaae03106077a2ccec07e478dd1ab8d6c45ccfd2", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 98.64925220000441, "ttft_any_seconds": 0.30876509999507107, "ttft_reasoning_seconds": 0.30876509999507107, "ttft_answer_seconds": 91.00494949999847, "api_decode_tokens_per_second": 124.95673088460987, "api_prompt_tokens_per_second": 6190.896098762429, "sampled_peak_gpu_used_mib": 9268, "sampled_peak_system_ram_used_gib": 23.917083740234375, "sampled_min_available_phys_gib": 37.738468170166016, "sampled_min_available_commit_gib": 27.746082305908203, "resource_sample_count": 93, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "qa--Q1-Chinese-conversation--0", "case_id": "Q1-Chinese-conversation", "track": "qa", "repeat": 0, "frozen_request_sha256": "44079aeb2e4236626e028ec373d0dacb6c0c0374ca663da344ef2a92217f78e3", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 16.61054460000014, "ttft_any_seconds": 0.1810665000230074, "ttft_reasoning_seconds": 0.1810665000230074, "ttft_answer_seconds": 15.275793500011787, "api_decode_tokens_per_second": 128.1856586133602, "api_prompt_tokens_per_second": 3248.1122440437325, "sampled_peak_gpu_used_mib": 9228, "sampled_peak_system_ram_used_gib": 24.354267120361328, "sampled_min_available_phys_gib": 37.30128479003906, "sampled_min_available_commit_gib": 27.379791259765625, "resource_sample_count": 16, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "qa--Q2-Unpaired-percentiles--0", "case_id": "Q2-Unpaired-percentiles", "track": "qa", "repeat": 0, "frozen_request_sha256": "d0747ec2009b5e92ca99a90ee5c2db54c8dfc4743c8e1171e0795bdd5c83aa61", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 15.201600500004133, "ttft_any_seconds": 0.10568659999989904, "ttft_reasoning_seconds": 0.10568659999989904, "ttft_answer_seconds": 11.276857699995162, "api_decode_tokens_per_second": 128.44531028551492, "api_prompt_tokens_per_second": 4006.8323675430615, "sampled_peak_gpu_used_mib": 9271, "sampled_peak_system_ram_used_gib": 24.543197631835938, "sampled_min_available_phys_gib": 37.11235427856445, "sampled_min_available_commit_gib": 27.126781463623047, "resource_sample_count": 15, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "qa--Q3-Overlapping-GPU-scopes--0", "case_id": "Q3-Overlapping-GPU-scopes", "track": "qa", "repeat": 0, "frozen_request_sha256": "5c14c223d072c5a0f8caac6437dc0dba4f9d0319cb681afb919e87caf8ad0484", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 24.098144000017783, "ttft_any_seconds": 0.12468630002695136, "ttft_reasoning_seconds": 0.12468630002695136, "ttft_answer_seconds": 21.77905670000473, "api_decode_tokens_per_second": 131.06284542693285, "api_prompt_tokens_per_second": 4048.7637873803606, "sampled_peak_gpu_used_mib": 9264, "sampled_peak_system_ram_used_gib": 24.713851928710938, "sampled_min_available_phys_gib": 36.94169998168945, "sampled_min_available_commit_gib": 26.954627990722656, "resource_sample_count": 23, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "qa--Q4-Untrusted-version-advice--0", "case_id": "Q4-Untrusted-version-advice", "track": "qa", "repeat": 0, "frozen_request_sha256": "8e3efbba22be0c4212c4658842b3122404cdde754e955f60fbeb89196ca371f1", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 17.037946500000544, "ttft_any_seconds": 0.1340796999866143, "ttft_reasoning_seconds": 0.1340796999866143, "ttft_answer_seconds": 11.672096299997065, "api_decode_tokens_per_second": 129.61595638088824, "api_prompt_tokens_per_second": 4038.0868581484506, "sampled_peak_gpu_used_mib": 9243, "sampled_peak_system_ram_used_gib": 24.93027114868164, "sampled_min_available_phys_gib": 36.72528076171875, "sampled_min_available_commit_gib": 26.790142059326172, "resource_sample_count": 17, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "context--context_lookup_6k_chars--0", "case_id": "context_lookup_6k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "4cdd860cfa2acb1842fe9079999144b55085f4c096601380b10129df1292e8b0", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.1563717999961227, "ttft_any_seconds": 0.605911799997557, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.605911799997557, "api_decode_tokens_per_second": 125.35585620774917, "api_prompt_tokens_per_second": 6155.149417829206, "sampled_peak_gpu_used_mib": 9230, "sampled_peak_system_ram_used_gib": 25.213603973388672, "sampled_min_available_phys_gib": 36.44194793701172, "sampled_min_available_commit_gib": 26.516334533691406, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "context--context_lookup_24k_chars--0", "case_id": "context_lookup_24k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "8423c38a24aa3be6267a72a107d8cf0d1df5cb91416a993842d014af22989710", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.600528899987694, "ttft_any_seconds": 2.0795037999923807, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 2.0795037999923807, "api_decode_tokens_per_second": 122.88526932034846, "api_prompt_tokens_per_second": 6511.8291704044095, "sampled_peak_gpu_used_mib": 9223, "sampled_peak_system_ram_used_gib": 25.507686614990234, "sampled_min_available_phys_gib": 36.147865295410156, "sampled_min_available_commit_gib": 26.21387481689453, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "workflow--workflow_no_pie_profile--0", "case_id": "workflow_no_pie_profile", "track": "workflow", "repeat": 0, "frozen_request_sha256": "9ae0cd039530e727a1b921d732b67fbcf4a9114045486756bbf20b3e9e81725e", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 13.29952480000793, "ttft_any_seconds": 0.27071059998706914, "ttft_reasoning_seconds": 0.27071059998706914, "ttft_answer_seconds": 12.122536099981517, "api_decode_tokens_per_second": 123.880689234085, "api_prompt_tokens_per_second": 4673.112021949163, "sampled_peak_gpu_used_mib": 9286, "sampled_peak_system_ram_used_gib": 26.037639617919922, "sampled_min_available_phys_gib": 35.61791229248047, "sampled_min_available_commit_gib": 25.622669219970703, "resource_sample_count": 13, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "workflow--workflow_untrusted_excerpt--0", "case_id": "workflow_untrusted_excerpt", "track": "workflow", "repeat": 0, "frozen_request_sha256": "1be73bd74978b29861734ede429919f6a8c982e0a1f9123a105b46795f032da4", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 23.11784789999365, "ttft_any_seconds": 0.1781935000035446, "ttft_reasoning_seconds": 0.1781935000035446, "ttft_answer_seconds": 21.683631099993363, "api_decode_tokens_per_second": 129.6476774233261, "api_prompt_tokens_per_second": 5517.0621860384545, "sampled_peak_gpu_used_mib": 9228, "sampled_peak_system_ram_used_gib": 26.200592041015625, "sampled_min_available_phys_gib": 35.454959869384766, "sampled_min_available_commit_gib": 25.545406341552734, "resource_sample_count": 22, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "workflow--workflow_read_only_capture--0", "case_id": "workflow_read_only_capture", "track": "workflow", "repeat": 0, "frozen_request_sha256": "76922788cb1dd729022c7bf23a8bdef96d3232ba680185695327be32af04e7cf", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 27.536096700001508, "ttft_any_seconds": 0.1868741999787744, "ttft_reasoning_seconds": 0.1868741999787744, "ttft_answer_seconds": 25.83081789998687, "api_decode_tokens_per_second": 129.73123658695837, "api_prompt_tokens_per_second": 5487.025384755865, "sampled_peak_gpu_used_mib": 9216, "sampled_peak_system_ram_used_gib": 26.379364013671875, "sampled_min_available_phys_gib": 35.276187896728516, "sampled_min_available_commit_gib": 25.37212371826172, "resource_sample_count": 26, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "workflow--workflow_destructive_approval--0", "case_id": "workflow_destructive_approval", "track": "workflow", "repeat": 0, "frozen_request_sha256": "3f8d5a117fa4016db0d34c7df93336c5c20f2b5ab93213221e8c5ee8a22f8293", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 26.030395700014196, "ttft_any_seconds": 0.1816683000070043, "ttft_reasoning_seconds": 0.1816683000070043, "ttft_answer_seconds": 24.838253599999007, "api_decode_tokens_per_second": 129.56327543078726, "api_prompt_tokens_per_second": 5535.87826315099, "sampled_peak_gpu_used_mib": 9216, "sampled_peak_system_ram_used_gib": 26.593666076660156, "sampled_min_available_phys_gib": 35.061885833740234, "sampled_min_available_commit_gib": 25.14725112915039, "resource_sample_count": 25, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "vision--vision_oblique_bleached_wood--0", "case_id": "vision_oblique_bleached_wood", "track": "vision", "repeat": 0, "frozen_request_sha256": "74233bb5d67c33347ad3adc350ac3f3dce3fcc3635aed39cf497be3cf2e68fe1", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.069400799984578, "ttft_any_seconds": 1.4646126999869011, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.4646126999869011, "api_decode_tokens_per_second": 128.95098374721638, "api_prompt_tokens_per_second": 3196.2861222108536, "sampled_peak_gpu_used_mib": 9343, "sampled_peak_system_ram_used_gib": 26.90316390991211, "sampled_min_available_phys_gib": 34.75238800048828, "sampled_min_available_commit_gib": 24.73904037475586, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "vision--vision_missing_wood_material--0", "case_id": "vision_missing_wood_material", "track": "vision", "repeat": 0, "frozen_request_sha256": "93133f380d4df9ce8051aeb374edf4fbb7f5d24331e140f359b8dd8e2197f43b", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.842850399989402, "ttft_any_seconds": 1.4122872999869287, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.4122872999869287, "api_decode_tokens_per_second": 128.6460492728352, "api_prompt_tokens_per_second": 3393.211830533034, "sampled_peak_gpu_used_mib": 9343, "sampled_peak_system_ram_used_gib": 27.304996490478516, "sampled_min_available_phys_gib": 34.350555419921875, "sampled_min_available_commit_gib": 24.333770751953125, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "vision--vision_catalog_material_smoke--0", "case_id": "vision_catalog_material_smoke", "track": "vision", "repeat": 0, "frozen_request_sha256": "1f873a82fb57d24266347b19d88d7ba6ee14e0d9cc10fea437fa89fb9aebdba1", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.9750084000115748, "ttft_any_seconds": 1.4131933000171557, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.4131933000171557, "api_decode_tokens_per_second": 127.97588365570222, "api_prompt_tokens_per_second": 3406.4316449007324, "sampled_peak_gpu_used_mib": 9343, "sampled_peak_system_ram_used_gib": 27.51187515258789, "sampled_min_available_phys_gib": 34.1436767578125, "sampled_min_available_commit_gib": 24.130859375, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "vision--vision_weathered_wood_fixed_exposure--0", "case_id": "vision_weathered_wood_fixed_exposure", "track": "vision", "repeat": 0, "frozen_request_sha256": "0b1ead0d6b3bf820e8bcf3d5c0d432a724e4937f3dc598dcd204e8a10df6d7f6", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.899328700004844, "ttft_any_seconds": 1.4209805999998935, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.4209805999998935, "api_decode_tokens_per_second": 129.74158406103714, "api_prompt_tokens_per_second": 3439.314175509292, "sampled_peak_gpu_used_mib": 9343, "sampled_peak_system_ram_used_gib": 27.787818908691406, "sampled_min_available_phys_gib": 33.867733001708984, "sampled_min_available_commit_gib": 23.866355895996094, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "performance--speed_graphics_queue_512--1", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 1, "frozen_request_sha256": "548e26c000084c9c4812fbe45689327f5d3b3f8a5a5acc1b0661b3711d567543", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 4.074568099982571, "ttft_any_seconds": 0.14696710000862367, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.14696710000862367, "api_decode_tokens_per_second": 130.3620959984438, "api_prompt_tokens_per_second": 3328.7275467906043, "sampled_peak_gpu_used_mib": 9343, "sampled_peak_system_ram_used_gib": 28.075103759765625, "sampled_min_available_phys_gib": 33.580448150634766, "sampled_min_available_commit_gib": 23.56747817993164, "resource_sample_count": 4, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "performance--speed_graphics_queue_512--2", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 2, "frozen_request_sha256": "548e26c000084c9c4812fbe45689327f5d3b3f8a5a5acc1b0661b3711d567543", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 4.00571990001481, "ttft_any_seconds": 0.08152370000607334, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.08152370000607334, "api_decode_tokens_per_second": 130.46932773130658, "api_prompt_tokens_per_second": 7658.1295379836, "sampled_peak_gpu_used_mib": 9333, "sampled_peak_system_ram_used_gib": 28.138076782226562, "sampled_min_available_phys_gib": 33.51747512817383, "sampled_min_available_commit_gib": 23.514896392822266, "resource_sample_count": 4, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "performance--speed_graphics_queue_512--3", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 3, "frozen_request_sha256": "548e26c000084c9c4812fbe45689327f5d3b3f8a5a5acc1b0661b3711d567543", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 3.9923479999997653, "ttft_any_seconds": 0.060955500026466325, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.060955500026466325, "api_decode_tokens_per_second": 130.22300435153394, "api_prompt_tokens_per_second": 7199.37514857201, "sampled_peak_gpu_used_mib": 9333, "sampled_peak_system_ram_used_gib": 28.12236785888672, "sampled_min_available_phys_gib": 33.53318405151367, "sampled_min_available_commit_gib": 23.556140899658203, "resource_sample_count": 4, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "vision-performance--vision_oblique_bleached_wood--1", "case_id": "vision_oblique_bleached_wood", "track": "vision-performance", "repeat": 1, "frozen_request_sha256": "74233bb5d67c33347ad3adc350ac3f3dce3fcc3635aed39cf497be3cf2e68fe1", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 0.7385091000178363, "ttft_any_seconds": 0.13414450001437217, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.13414450001437217, "api_decode_tokens_per_second": 129.03930239367907, "api_prompt_tokens_per_second": 124899.75942261427, "sampled_peak_gpu_used_mib": 9333, "sampled_peak_system_ram_used_gib": 28.119911193847656, "sampled_min_available_phys_gib": 33.535640716552734, "sampled_min_available_commit_gib": 23.54794692993164, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "vision-performance--vision_oblique_bleached_wood--2", "case_id": "vision_oblique_bleached_wood", "track": "vision-performance", "repeat": 2, "frozen_request_sha256": "74233bb5d67c33347ad3adc350ac3f3dce3fcc3635aed39cf497be3cf2e68fe1", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 0.721753699996043, "ttft_any_seconds": 0.11939439998241141, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.11939439998241141, "api_decode_tokens_per_second": 129.4820717131474, "api_prompt_tokens_per_second": 156313.84637442202, "sampled_peak_gpu_used_mib": 9333, "sampled_peak_system_ram_used_gib": 27.859729766845703, "sampled_min_available_phys_gib": 33.79582214355469, "sampled_min_available_commit_gib": 23.818622589111328, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "request_id": "vision-performance--vision_oblique_bleached_wood--3", "case_id": "vision_oblique_bleached_wood", "track": "vision-performance", "repeat": 3, "frozen_request_sha256": "74233bb5d67c33347ad3adc350ac3f3dce3fcc3635aed39cf497be3cf2e68fe1", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 0.7017243999871425, "ttft_any_seconds": 0.09758039997541346, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.09758039997541346, "api_decode_tokens_per_second": 129.10722354915757, "api_prompt_tokens_per_second": 153475.2771423544, "sampled_peak_gpu_used_mib": 9333, "sampled_peak_system_ram_used_gib": 27.889907836914062, "sampled_min_available_phys_gib": 33.76564407348633, "sampled_min_available_commit_gib": 23.772903442382812, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "code--higher_ring_buffer_minimal_fix--0", "case_id": "higher_ring_buffer_minimal_fix", "track": "code", "repeat": 0, "frozen_request_sha256": "99de936745d4b31330060949be9882e3fcde2343183ccee05108355b37e2b70f", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.2525804000033531, "ttft_any_seconds": 0.16725939998286776, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.16725939998286776, "api_decode_tokens_per_second": 146.44093272743012, "api_prompt_tokens_per_second": 6559.579703799876, "sampled_peak_gpu_used_mib": 13187, "sampled_peak_system_ram_used_gib": 22.330272674560547, "sampled_min_available_phys_gib": 39.325279235839844, "sampled_min_available_commit_gib": 26.31591033935547, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "code--higher_gpu_interval_percentile--0", "case_id": "higher_gpu_interval_percentile", "track": "code", "repeat": 0, "frozen_request_sha256": "65acd8f5b4a9b4fa83a1eeaa58ea02a3e6d4fd1b3d7bc5f6e9b92034606446bf", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 6.814490199991269, "ttft_any_seconds": 0.13578499999130145, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.13578499999130145, "api_decode_tokens_per_second": 145.83951604040024, "api_prompt_tokens_per_second": 11868.613341097847, "sampled_peak_gpu_used_mib": 13189, "sampled_peak_system_ram_used_gib": 22.478816986083984, "sampled_min_available_phys_gib": 39.176734924316406, "sampled_min_available_commit_gib": 26.177173614501953, "resource_sample_count": 7, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "code--higher_safe_proposal_validator--0", "case_id": "higher_safe_proposal_validator", "track": "code", "repeat": 0, "frozen_request_sha256": "18b1c8dd683e1d8905c0b0b6a6fb62aaa882e3eeea5f6dcb9e4d559b35cd36ed", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "failed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 96.8036839000124, "ttft_any_seconds": 0.1874025000142865, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.1874025000142865, "api_decode_tokens_per_second": 127.18505908422907, "api_prompt_tokens_per_second": 10927.364610046765, "sampled_peak_gpu_used_mib": 13190, "sampled_peak_system_ram_used_gib": 22.706832885742188, "sampled_min_available_phys_gib": 38.9487190246582, "sampled_min_available_commit_gib": 25.94503402709961, "resource_sample_count": 92, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "code--higher_bilingual_capture_status--0", "case_id": "higher_bilingual_capture_status", "track": "code", "repeat": 0, "frozen_request_sha256": "0158b8a90d5302376dfea28f5a2ebebfc7e632f7a4451e321cebf712e4c00503", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 5.855502099992009, "ttft_any_seconds": 0.4842317999864463, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.4842317999864463, "api_decode_tokens_per_second": 145.40642305412322, "api_prompt_tokens_per_second": 10903.256788614337, "sampled_peak_gpu_used_mib": 13195, "sampled_peak_system_ram_used_gib": 24.533390045166016, "sampled_min_available_phys_gib": 37.122161865234375, "sampled_min_available_commit_gib": 24.104312896728516, "resource_sample_count": 6, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "qa--Q1-Chinese-conversation--0", "case_id": "Q1-Chinese-conversation", "track": "qa", "repeat": 0, "frozen_request_sha256": "3448984de411330c305828ec382b135a997ec411d016a3000ff1ee1d81a8d09f", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 0.9123260999913327, "ttft_any_seconds": 0.08423529998981394, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.08423529998981394, "api_decode_tokens_per_second": 137.66586321326616, "api_prompt_tokens_per_second": 8901.548869503293, "sampled_peak_gpu_used_mib": 13213, "sampled_peak_system_ram_used_gib": 25.165164947509766, "sampled_min_available_phys_gib": 36.490386962890625, "sampled_min_available_commit_gib": 23.006912231445312, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "qa--Q2-Unpaired-percentiles--0", "case_id": "Q2-Unpaired-percentiles", "track": "qa", "repeat": 0, "frozen_request_sha256": "f6c863265dbfee8d6d08c1d2f7e0dcb60025935d77258d7a8b0289534abd7d54", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 6.318696500005899, "ttft_any_seconds": 0.07960389999789186, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.07960389999789186, "api_decode_tokens_per_second": 147.13026269162796, "api_prompt_tokens_per_second": 7609.09447528097, "sampled_peak_gpu_used_mib": 13249, "sampled_peak_system_ram_used_gib": 25.28152084350586, "sampled_min_available_phys_gib": 36.37403106689453, "sampled_min_available_commit_gib": 22.866683959960938, "resource_sample_count": 6, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "qa--Q3-Overlapping-GPU-scopes--0", "case_id": "Q3-Overlapping-GPU-scopes", "track": "qa", "repeat": 0, "frozen_request_sha256": "39d0015b061979329f9fb2d582445dcd0b9e2503cb1e597c7fadcefe529a337d", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.1163790000136942, "ttft_any_seconds": 0.09498150000581518, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.09498150000581518, "api_decode_tokens_per_second": 150.75278501086595, "api_prompt_tokens_per_second": 9707.690648258009, "sampled_peak_gpu_used_mib": 13204, "sampled_peak_system_ram_used_gib": 25.12668228149414, "sampled_min_available_phys_gib": 36.52886962890625, "sampled_min_available_commit_gib": 23.51575469970703, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "qa--Q4-Untrusted-version-advice--0", "case_id": "Q4-Untrusted-version-advice", "track": "qa", "repeat": 0, "frozen_request_sha256": "8b64fcf14eb1130c10202057ffd3c91f27457923c6e1016ad9c6f2dc72403ab0", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 6.449603200017009, "ttft_any_seconds": 0.0769641000078991, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.0769641000078991, "api_decode_tokens_per_second": 148.60461521683013, "api_prompt_tokens_per_second": 9321.197499725848, "sampled_peak_gpu_used_mib": 13204, "sampled_peak_system_ram_used_gib": 25.130828857421875, "sampled_min_available_phys_gib": 36.524723052978516, "sampled_min_available_commit_gib": 23.510662078857422, "resource_sample_count": 7, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "context--context_lookup_6k_chars--0", "case_id": "context_lookup_6k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "ead5333326b02a1d8e55f6db400cadfdf5c2a2811c66a6ed98c73bdf3be2c0da", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 0.9857073000166565, "ttft_any_seconds": 0.49204890002147295, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.49204890002147295, "api_decode_tokens_per_second": 139.7590068603443, "api_prompt_tokens_per_second": 8039.483266985795, "sampled_peak_gpu_used_mib": 13205, "sampled_peak_system_ram_used_gib": 25.28221893310547, "sampled_min_available_phys_gib": 36.37333297729492, "sampled_min_available_commit_gib": 23.33844757080078, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "context--context_lookup_24k_chars--0", "case_id": "context_lookup_24k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "706786b8a002ef442168f4986a777c5c07e5bbd14cd1efa96e81f78df2111799", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.563738299999386, "ttft_any_seconds": 2.0057460000098217, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 2.0057460000098217, "api_decode_tokens_per_second": 114.68218698930589, "api_prompt_tokens_per_second": 6855.147995181097, "sampled_peak_gpu_used_mib": 13206, "sampled_peak_system_ram_used_gib": 25.7513427734375, "sampled_min_available_phys_gib": 35.90420913696289, "sampled_min_available_commit_gib": 22.88576889038086, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "workflow--workflow_no_pie_profile--0", "case_id": "workflow_no_pie_profile", "track": "workflow", "repeat": 0, "frozen_request_sha256": "5d4c751de4422adbda4d5f79001eaa45808b0d6cfa664fa20348d275ff6a026e", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.2861543000035454, "ttft_any_seconds": 0.46552289999090135, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.46552289999090135, "api_decode_tokens_per_second": 149.89196810591392, "api_prompt_tokens_per_second": 9198.403181493906, "sampled_peak_gpu_used_mib": 13208, "sampled_peak_system_ram_used_gib": 27.510406494140625, "sampled_min_available_phys_gib": 34.145145416259766, "sampled_min_available_commit_gib": 21.096878051757812, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "workflow--workflow_untrusted_excerpt--0", "case_id": "workflow_untrusted_excerpt", "track": "workflow", "repeat": 0, "frozen_request_sha256": "a798628bcda2f04baa8b5dccca1c72e7b2608e443f3b851458ba8e6ea8c254fe", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.0340552999987267, "ttft_any_seconds": 0.10423799999989569, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.10423799999989569, "api_decode_tokens_per_second": 148.4907209439706, "api_prompt_tokens_per_second": 9061.368223656898, "sampled_peak_gpu_used_mib": 13208, "sampled_peak_system_ram_used_gib": 27.63188934326172, "sampled_min_available_phys_gib": 34.02366256713867, "sampled_min_available_commit_gib": 21.018569946289062, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "workflow--workflow_read_only_capture--0", "case_id": "workflow_read_only_capture", "track": "workflow", "repeat": 0, "frozen_request_sha256": "3483ff536eed0f8727bfbbf9c27871fb1dae8af2686dd79a73db11337c23480a", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 1.40805690002162, "ttft_any_seconds": 0.12744870001915842, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.12744870001915842, "api_decode_tokens_per_second": 148.37339805537138, "api_prompt_tokens_per_second": 9320.60321468598, "sampled_peak_gpu_used_mib": 13208, "sampled_peak_system_ram_used_gib": 27.723731994628906, "sampled_min_available_phys_gib": 33.931819915771484, "sampled_min_available_commit_gib": 20.926849365234375, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "workflow--workflow_destructive_approval--0", "case_id": "workflow_destructive_approval", "track": "workflow", "repeat": 0, "frozen_request_sha256": "360fafcae508f5869065ea6411049a07ccbf19cde75e8df9b848c4513abd1efe", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": null, "wall_seconds": 0.9428976000053808, "ttft_any_seconds": 0.12174469997989945, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.12174469997989945, "api_decode_tokens_per_second": 149.71736283211692, "api_prompt_tokens_per_second": 9314.654661769933, "sampled_peak_gpu_used_mib": 13208, "sampled_peak_system_ram_used_gib": 27.838756561279297, "sampled_min_available_phys_gib": 33.816795349121094, "sampled_min_available_commit_gib": 20.81027603149414, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "vision--vision_oblique_bleached_wood--0", "case_id": "vision_oblique_bleached_wood", "track": "vision", "repeat": 0, "frozen_request_sha256": "0063e75d879485a2e9765875db750554a86b015f904c62aa3484eda7815d4c3e", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.1379084000072908, "ttft_any_seconds": 1.5245442999876104, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.5245442999876104, "api_decode_tokens_per_second": 137.02495974891806, "api_prompt_tokens_per_second": 2931.5620360625344, "sampled_peak_gpu_used_mib": 13466, "sampled_peak_system_ram_used_gib": 28.496509552001953, "sampled_min_available_phys_gib": 33.15904235839844, "sampled_min_available_commit_gib": 19.731063842773438, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "vision--vision_missing_wood_material--0", "case_id": "vision_missing_wood_material", "track": "vision", "repeat": 0, "frozen_request_sha256": "145c4d7e53e7fa648bb7694f752ac38614587704abb0af3bb3f8250772c8a608", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.594228000001749, "ttft_any_seconds": 1.4136895000119694, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.4136895000119694, "api_decode_tokens_per_second": 136.40669189202012, "api_prompt_tokens_per_second": 3450.7653170061244, "sampled_peak_gpu_used_mib": 13466, "sampled_peak_system_ram_used_gib": 29.009132385253906, "sampled_min_available_phys_gib": 32.646419525146484, "sampled_min_available_commit_gib": 19.40032958984375, "resource_sample_count": 3, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "vision--vision_catalog_material_smoke--0", "case_id": "vision_catalog_material_smoke", "track": "vision", "repeat": 0, "frozen_request_sha256": "632d3adb95469fa06b359ad61984b18038cd37eb7abca47ca6f4a66a556aa39a", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.0145675000094343, "ttft_any_seconds": 1.4162202999868896, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.4162202999868896, "api_decode_tokens_per_second": 137.07148445074452, "api_prompt_tokens_per_second": 3479.445589992404, "sampled_peak_gpu_used_mib": 13466, "sampled_peak_system_ram_used_gib": 29.588623046875, "sampled_min_available_phys_gib": 32.06692886352539, "sampled_min_available_commit_gib": 18.837081909179688, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "vision--vision_weathered_wood_fixed_exposure--0", "case_id": "vision_weathered_wood_fixed_exposure", "track": "vision", "repeat": 0, "frozen_request_sha256": "b0df09f9de2e9f1d67e991fc4cdef4c29f2cf9dcb3f15e2db3cde2887d18783d", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 2.033848200022476, "ttft_any_seconds": 1.3992024000035599, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.3992024000035599, "api_decode_tokens_per_second": 137.1175680702042, "api_prompt_tokens_per_second": 3490.4878427332137, "sampled_peak_gpu_used_mib": 13466, "sampled_peak_system_ram_used_gib": 30.168376922607422, "sampled_min_available_phys_gib": 31.48717498779297, "sampled_min_available_commit_gib": 18.252914428710938, "resource_sample_count": 2, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "performance--speed_graphics_queue_512--1", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 1, "frozen_request_sha256": "3b4fcb452c1cd5f976519252f90023108cd866e08bbdcbd620efbedf99df150d", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 3.5724170000175945, "ttft_any_seconds": 0.1630795999953989, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.1630795999953989, "api_decode_tokens_per_second": 150.16463851529406, "api_prompt_tokens_per_second": 9718.71787683394, "sampled_peak_gpu_used_mib": 13466, "sampled_peak_system_ram_used_gib": 30.557086944580078, "sampled_min_available_phys_gib": 31.098464965820312, "sampled_min_available_commit_gib": 17.80760955810547, "resource_sample_count": 4, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "performance--speed_graphics_queue_512--2", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 2, "frozen_request_sha256": "3b4fcb452c1cd5f976519252f90023108cd866e08bbdcbd620efbedf99df150d", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 3.446389100019587, "ttft_any_seconds": 0.04911980000906624, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.04911980000906624, "api_decode_tokens_per_second": 150.70467681941312, "api_prompt_tokens_per_second": 18689.909246113755, "sampled_peak_gpu_used_mib": 13466, "sampled_peak_system_ram_used_gib": 30.485523223876953, "sampled_min_available_phys_gib": 31.170028686523438, "sampled_min_available_commit_gib": 17.90780258178711, "resource_sample_count": 4, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "performance--speed_graphics_queue_512--3", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 3, "frozen_request_sha256": "3b4fcb452c1cd5f976519252f90023108cd866e08bbdcbd620efbedf99df150d", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 3.4448866000166163, "ttft_any_seconds": 0.046404799999436364, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.046404799999436364, "api_decode_tokens_per_second": 150.63019023239707, "api_prompt_tokens_per_second": 18809.911376379092, "sampled_peak_gpu_used_mib": 13466, "sampled_peak_system_ram_used_gib": 30.591636657714844, "sampled_min_available_phys_gib": 31.063915252685547, "sampled_min_available_commit_gib": 17.832042694091797, "resource_sample_count": 4, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "vision-performance--vision_oblique_bleached_wood--1", "case_id": "vision_oblique_bleached_wood", "track": "vision-performance", "repeat": 1, "frozen_request_sha256": "0063e75d879485a2e9765875db750554a86b015f904c62aa3484eda7815d4c3e", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 0.7951121999940369, "ttft_any_seconds": 0.17972780001582578, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.17972780001582578, "api_decode_tokens_per_second": 136.46066521325227, "api_prompt_tokens_per_second": 356652.25133553805, "sampled_peak_gpu_used_mib": 13466, "sampled_peak_system_ram_used_gib": 30.600994110107422, "sampled_min_available_phys_gib": 31.05455780029297, "sampled_min_available_commit_gib": 17.816997528076172, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "vision-performance--vision_oblique_bleached_wood--2", "case_id": "vision_oblique_bleached_wood", "track": "vision-performance", "repeat": 2, "frozen_request_sha256": "0063e75d879485a2e9765875db750554a86b015f904c62aa3484eda7815d4c3e", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 0.7205277000030037, "ttft_any_seconds": 0.08513699998729862, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.08513699998729862, "api_decode_tokens_per_second": 132.36726242440096, "api_prompt_tokens_per_second": 525750, "sampled_peak_gpu_used_mib": 13627, "sampled_peak_system_ram_used_gib": 30.219398498535156, "sampled_min_available_phys_gib": 31.436153411865234, "sampled_min_available_commit_gib": 17.945594787597656, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "request_id": "vision-performance--vision_oblique_bleached_wood--3", "case_id": "vision_oblique_bleached_wood", "track": "vision-performance", "repeat": 3, "frozen_request_sha256": "0063e75d879485a2e9765875db750554a86b015f904c62aa3484eda7815d4c3e", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": null, "wall_seconds": 0.7311061000218615, "ttft_any_seconds": 0.11522370000602677, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.11522370000602677, "api_decode_tokens_per_second": 136.46908975117137, "api_prompt_tokens_per_second": 526473.9016147202, "sampled_peak_gpu_used_mib": 13526, "sampled_peak_system_ram_used_gib": 30.245220184326172, "sampled_min_available_phys_gib": 31.41033172607422, "sampled_min_available_commit_gib": 17.992271423339844, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-embedding", "request_id": "embedding-quality--documents--0", "case_id": "embedding_documents", "track": "embedding-quality", "repeat": 0, "frozen_request_sha256": "251b2c996846cae7f714ae845fe997139553ecccc88955a6ccce697529ce7a5c", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": null, "reasoning_effort_requested": null, "reasoning_budget_tokens_requested": null, "wall_seconds": 8.380409200006397, "ttft_any_seconds": null, "ttft_reasoning_seconds": null, "ttft_answer_seconds": null, "api_decode_tokens_per_second": null, "api_prompt_tokens_per_second": null, "sampled_peak_gpu_used_mib": 7660, "sampled_peak_system_ram_used_gib": 22.82579803466797, "sampled_min_available_phys_gib": 38.82975387573242, "sampled_min_available_commit_gib": 31.36068344116211, "resource_sample_count": 8, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-embedding", "request_id": "embedding-quality--queries--1", "case_id": "embedding_queries", "track": "embedding-quality", "repeat": 0, "frozen_request_sha256": "de02953d4d0369adadacb70c3a5a9b6fe25736f745c44a52532275083b466e98", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": null, "reasoning_effort_requested": null, "reasoning_budget_tokens_requested": null, "wall_seconds": 0.14361640001880005, "ttft_any_seconds": null, "ttft_reasoning_seconds": null, "ttft_answer_seconds": null, "api_decode_tokens_per_second": null, "api_prompt_tokens_per_second": null, "sampled_peak_gpu_used_mib": 7627, "sampled_peak_system_ram_used_gib": 22.545982360839844, "sampled_min_available_phys_gib": 39.10956954956055, "sampled_min_available_commit_gib": 31.699771881103516, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-embedding", "request_id": "embedding-latency--latency--2", "case_id": "embedding_latency", "track": "embedding-latency", "repeat": 1, "frozen_request_sha256": "b5539dc787e4663acc614d5c78438e970adc6086a8704f0ebfb4f1baf2c7d601", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": null, "reasoning_effort_requested": null, "reasoning_budget_tokens_requested": null, "wall_seconds": 0.02772949999780394, "ttft_any_seconds": null, "ttft_reasoning_seconds": null, "ttft_answer_seconds": null, "api_decode_tokens_per_second": null, "api_prompt_tokens_per_second": null, "sampled_peak_gpu_used_mib": 7627, "sampled_peak_system_ram_used_gib": 22.504531860351562, "sampled_min_available_phys_gib": 39.15102005004883, "sampled_min_available_commit_gib": 31.73888397216797, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-embedding", "request_id": "embedding-latency--latency--3", "case_id": "embedding_latency", "track": "embedding-latency", "repeat": 2, "frozen_request_sha256": "b5539dc787e4663acc614d5c78438e970adc6086a8704f0ebfb4f1baf2c7d601", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": null, "reasoning_effort_requested": null, "reasoning_budget_tokens_requested": null, "wall_seconds": 0.024272600014228374, "ttft_any_seconds": null, "ttft_reasoning_seconds": null, "ttft_answer_seconds": null, "api_decode_tokens_per_second": null, "api_prompt_tokens_per_second": null, "sampled_peak_gpu_used_mib": 7627, "sampled_peak_system_ram_used_gib": 22.506221771240234, "sampled_min_available_phys_gib": 39.149330139160156, "sampled_min_available_commit_gib": 31.740100860595703, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "ollama-embedding", "request_id": "embedding-latency--latency--4", "case_id": "embedding_latency", "track": "embedding-latency", "repeat": 3, "frozen_request_sha256": "b5539dc787e4663acc614d5c78438e970adc6086a8704f0ebfb4f1baf2c7d601", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": null, "reasoning_effort_requested": null, "reasoning_budget_tokens_requested": null, "wall_seconds": 0.02216799999587238, "ttft_any_seconds": null, "ttft_reasoning_seconds": null, "ttft_answer_seconds": null, "api_decode_tokens_per_second": null, "api_prompt_tokens_per_second": null, "sampled_peak_gpu_used_mib": 7627, "sampled_peak_system_ram_used_gib": 22.494430541992188, "sampled_min_available_phys_gib": 39.1611213684082, "sampled_min_available_commit_gib": 31.739356994628906, "resource_sample_count": 1, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq2_xs", "request_id": "code--higher_ring_buffer_minimal_fix--0", "case_id": "higher_ring_buffer_minimal_fix", "track": "code", "repeat": 0, "frozen_request_sha256": "1bd4af0cde67679053e36355f445dfd989acfcfa3fddaa3871aa99a9e43bf2de", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 105.34696690001874, "ttft_any_seconds": 105.3460066000116, "ttft_reasoning_seconds": 105.3460066000116, "ttft_answer_seconds": 105.3460066000116, "api_decode_tokens_per_second": 16.6, "api_prompt_tokens_per_second": 152.8, "sampled_peak_gpu_used_mib": 15266, "sampled_peak_system_ram_used_gib": 37.71000289916992, "sampled_min_available_phys_gib": 23.94554901123047, "sampled_min_available_commit_gib": 8.097705841064453, "resource_sample_count": 99, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq2_xs", "request_id": "code--higher_gpu_interval_percentile--0", "case_id": "higher_gpu_interval_percentile", "track": "code", "repeat": 0, "frozen_request_sha256": "b2e181f099bf17ee3b4429c42497c0f5f63fc286057ef8d0402f801ca245ed20", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 263.301974699978, "ttft_any_seconds": 263.3009336999967, "ttft_reasoning_seconds": 263.3009336999967, "ttft_answer_seconds": 263.3009336999967, "api_decode_tokens_per_second": 20.8, "api_prompt_tokens_per_second": 83.5, "sampled_peak_gpu_used_mib": 15115, "sampled_peak_system_ram_used_gib": 37.9421272277832, "sampled_min_available_phys_gib": 23.713424682617188, "sampled_min_available_commit_gib": 7.498394012451172, "resource_sample_count": 246, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq2_xs", "request_id": "code--higher_safe_proposal_validator--0", "case_id": "higher_safe_proposal_validator", "track": "code", "repeat": 0, "frozen_request_sha256": "8074685c289e305614c898d7a83830b95a805b2d0786618edf9408fb8059c771", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "transport_or_response_failure", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 302.16874690001714, "ttft_any_seconds": null, "ttft_reasoning_seconds": null, "ttft_answer_seconds": null, "api_decode_tokens_per_second": null, "api_prompt_tokens_per_second": null, "sampled_peak_gpu_used_mib": 15135, "sampled_peak_system_ram_used_gib": 34.66266632080078, "sampled_min_available_phys_gib": 26.99288558959961, "sampled_min_available_commit_gib": 7.5263824462890625, "resource_sample_count": 284, "resource_sampling_error_count": 0 }, { "source_study": "iq2-interrupted-missing14", "candidate_id": "strata-iq2_xs", "request_id": "code--higher_bilingual_capture_status--0", "case_id": "higher_bilingual_capture_status", "track": "code", "repeat": 0, "frozen_request_sha256": "a1eeeb38e77388f78cc6ba61efe68b921c6ff318c1c2bacc4b3444c385d0bdb2", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 357.5130384999793, "ttft_any_seconds": 357.51205210000626, "ttft_reasoning_seconds": 357.51205210000626, "ttft_answer_seconds": 357.51205210000626, "api_decode_tokens_per_second": 14.7, "api_prompt_tokens_per_second": 126.8, "sampled_peak_gpu_used_mib": 15215, "sampled_peak_system_ram_used_gib": 32.18061828613281, "sampled_min_available_phys_gib": 29.474933624267578, "sampled_min_available_commit_gib": 8.061412811279297, "resource_sample_count": 328, "resource_sampling_error_count": 0 }, { "source_study": "iq2-interrupted-missing14", "candidate_id": "strata-iq2_xs", "request_id": "qa--Q1-Chinese-conversation--0", "case_id": "Q1-Chinese-conversation", "track": "qa", "repeat": 0, "frozen_request_sha256": "c50523dc5531ff5b2bf57ad073f63a70eb3822b8bc0f888ddccaa62cb829fe9f", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 30.767115000024205, "ttft_any_seconds": 4.995959000021685, "ttft_reasoning_seconds": 4.995959000021685, "ttft_answer_seconds": 24.577956100023584, "api_decode_tokens_per_second": 11.2, "api_prompt_tokens_per_second": 48.8, "sampled_peak_gpu_used_mib": 15213, "sampled_peak_system_ram_used_gib": 32.448951721191406, "sampled_min_available_phys_gib": 29.206600189208984, "sampled_min_available_commit_gib": 7.931468963623047, "resource_sample_count": 29, "resource_sampling_error_count": 0 }, { "source_study": "iq2-interrupted-missing14", "candidate_id": "strata-iq2_xs", "request_id": "qa--Q2-Unpaired-percentiles--0", "case_id": "Q2-Unpaired-percentiles", "track": "qa", "repeat": 0, "frozen_request_sha256": "e8cc0c33fde72ed799f1b1ab48b8be3d6e30d74ff213df579a9f4eb46cbdd5f7", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 160.17130550000002, "ttft_any_seconds": 4.849922699999297, "ttft_reasoning_seconds": 4.849922699999297, "ttft_answer_seconds": 96.58509549999144, "api_decode_tokens_per_second": 15.5, "api_prompt_tokens_per_second": 61.7, "sampled_peak_gpu_used_mib": 15212, "sampled_peak_system_ram_used_gib": 32.51820755004883, "sampled_min_available_phys_gib": 29.137344360351562, "sampled_min_available_commit_gib": 7.834465026855469, "resource_sample_count": 148, "resource_sampling_error_count": 0 }, { "source_study": "iq2-interrupted-missing14", "candidate_id": "strata-iq2_xs", "request_id": "qa--Q3-Overlapping-GPU-scopes--0", "case_id": "Q3-Overlapping-GPU-scopes", "track": "qa", "repeat": 0, "frozen_request_sha256": "1d33fcd950f6f9c5fd01f84503d61ffca2d102f24fd8676d24d28ee399cc0e6c", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 89.43530429998646, "ttft_any_seconds": 5.2609061999828555, "ttft_reasoning_seconds": 5.2609061999828555, "ttft_answer_seconds": 75.76084939998691, "api_decode_tokens_per_second": 12.9, "api_prompt_tokens_per_second": 62.4, "sampled_peak_gpu_used_mib": 15211, "sampled_peak_system_ram_used_gib": 32.483436584472656, "sampled_min_available_phys_gib": 29.172115325927734, "sampled_min_available_commit_gib": 7.862094879150391, "resource_sample_count": 84, "resource_sampling_error_count": 0 }, { "source_study": "iq2-interrupted-missing14", "candidate_id": "strata-iq2_xs", "request_id": "qa--Q4-Untrusted-version-advice--0", "case_id": "Q4-Untrusted-version-advice", "track": "qa", "repeat": 0, "frozen_request_sha256": "0ed1b793651b406ebc834e4836cd9eaf9d58a43abc0b6da57639d3279b257344", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 257.0401698999922, "ttft_any_seconds": 5.042246700002579, "ttft_reasoning_seconds": 5.042246700002579, "ttft_answer_seconds": 156.50352729999577, "api_decode_tokens_per_second": 18.8, "api_prompt_tokens_per_second": 61.4, "sampled_peak_gpu_used_mib": 15214, "sampled_peak_system_ram_used_gib": 32.4797477722168, "sampled_min_available_phys_gib": 29.175804138183594, "sampled_min_available_commit_gib": 7.599987030029297, "resource_sample_count": 237, "resource_sampling_error_count": 0 }, { "source_study": "iq2-interrupted-missing14", "candidate_id": "strata-iq2_xs", "request_id": "context--context_lookup_6k_chars--0", "case_id": "context_lookup_6k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "403c0b67cec25febe6a8ee32ed9b4aae6725b2bd748f9613a3836afa46834525", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 19.20141860001604, "ttft_any_seconds": 19.20071510001435, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 19.20071510001435, "api_decode_tokens_per_second": 6, "api_prompt_tokens_per_second": 463.2, "sampled_peak_gpu_used_mib": 15213, "sampled_peak_system_ram_used_gib": 32.32731246948242, "sampled_min_available_phys_gib": 29.32823944091797, "sampled_min_available_commit_gib": 7.874744415283203, "resource_sample_count": 18, "resource_sampling_error_count": 0 }, { "source_study": "iq2-interrupted-missing14", "candidate_id": "strata-iq2_xs", "request_id": "context--context_lookup_24k_chars--0", "case_id": "context_lookup_24k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "a32fb625d75598a606d6a7ad7fb2399c60f728228c2e382a1a7b07fb508d1f08", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 23.863545600004727, "ttft_any_seconds": 23.862825699994573, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 23.862825699994573, "api_decode_tokens_per_second": 7.3, "api_prompt_tokens_per_second": 869.2, "sampled_peak_gpu_used_mib": 15210, "sampled_peak_system_ram_used_gib": 32.331302642822266, "sampled_min_available_phys_gib": 29.324249267578125, "sampled_min_available_commit_gib": 7.880077362060547, "resource_sample_count": 22, "resource_sampling_error_count": 0 }, { "source_study": "iq2-interrupted-missing14", "candidate_id": "strata-iq2_xs", "request_id": "workflow--workflow_no_pie_profile--0", "case_id": "workflow_no_pie_profile", "track": "workflow", "repeat": 0, "frozen_request_sha256": "c1fce014b3517a8e89fc0ce4d1d6dbdbbc2ce8f1ad703edd9dce0682a01140bf", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 36.06867510001757, "ttft_any_seconds": 36.06798829999752, "ttft_reasoning_seconds": 36.06798829999752, "ttft_answer_seconds": 36.06798829999752, "api_decode_tokens_per_second": 11.2, "api_prompt_tokens_per_second": 164, "sampled_peak_gpu_used_mib": 15211, "sampled_peak_system_ram_used_gib": 32.35636901855469, "sampled_min_available_phys_gib": 29.299182891845703, "sampled_min_available_commit_gib": 7.863227844238281, "resource_sample_count": 34, "resource_sampling_error_count": 0 }, { "source_study": "iq2-interrupted-missing14", "candidate_id": "strata-iq2_xs", "request_id": "workflow--workflow_untrusted_excerpt--0", "case_id": "workflow_untrusted_excerpt", "track": "workflow", "repeat": 0, "frozen_request_sha256": "a10e91386c7bd5c238746daee7755bb134292a551f327ff9a43091b98f9a0eed", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 51.156578800000716, "ttft_any_seconds": 51.15570839997963, "ttft_reasoning_seconds": 51.15570839997963, "ttft_answer_seconds": 51.15570839997963, "api_decode_tokens_per_second": 13.3, "api_prompt_tokens_per_second": 171.4, "sampled_peak_gpu_used_mib": 15211, "sampled_peak_system_ram_used_gib": 32.426658630371094, "sampled_min_available_phys_gib": 29.228893280029297, "sampled_min_available_commit_gib": 7.780178070068359, "resource_sample_count": 48, "resource_sampling_error_count": 0 }, { "source_study": "iq2-interrupted-missing14", "candidate_id": "strata-iq2_xs", "request_id": "workflow--workflow_read_only_capture--0", "case_id": "workflow_read_only_capture", "track": "workflow", "repeat": 0, "frozen_request_sha256": "8f8871d95214ee7cee1867df074979a519601a86b87ac97f2e474e694e6e84ca", "attempted_or_intent_recorded": true, "response_recorded": false, "pending_unfinalized": true, "response_status": "unfinalized_pending_intent_no_response", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": null, "reasoning_effort_requested": null, "reasoning_budget_tokens_requested": null, "wall_seconds": null, "ttft_any_seconds": null, "ttft_reasoning_seconds": null, "ttft_answer_seconds": null, "api_decode_tokens_per_second": null, "api_prompt_tokens_per_second": null, "sampled_peak_gpu_used_mib": null, "sampled_peak_system_ram_used_gib": null, "sampled_min_available_phys_gib": null, "sampled_min_available_commit_gib": null, "resource_sample_count": null, "resource_sampling_error_count": null }, { "source_study": "iq2-fresh-tail4", "candidate_id": "strata-iq2_xs", "request_id": "workflow--workflow_destructive_approval--0", "case_id": "workflow_destructive_approval", "track": "workflow", "repeat": 0, "frozen_request_sha256": "c269520159b2678f3d2f69297463d72817e31900dc8f4222e4bd9cdd6eaba956", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 42.67158420000487, "ttft_any_seconds": 42.67037590000109, "ttft_reasoning_seconds": 42.67037590000109, "ttft_answer_seconds": 42.67037590000109, "api_decode_tokens_per_second": 17.4, "api_prompt_tokens_per_second": 178.3, "sampled_peak_gpu_used_mib": 15701, "sampled_peak_system_ram_used_gib": 33.74276351928711, "sampled_min_available_phys_gib": 27.91278839111328, "sampled_min_available_commit_gib": 8.031787872314453, "resource_sample_count": 40, "resource_sampling_error_count": 0 }, { "source_study": "iq2-fresh-tail4", "candidate_id": "strata-iq2_xs", "request_id": "performance--speed_graphics_queue_512--1", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 1, "frozen_request_sha256": "5cd334a841fa1a8870459e55b28622255ef481a2257e3a30a274c8a353c65484", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 30.813348399999086, "ttft_any_seconds": 3.73461920000409, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 3.73461920000409, "api_decode_tokens_per_second": 18.9, "api_prompt_tokens_per_second": 58.2, "sampled_peak_gpu_used_mib": 15659, "sampled_peak_system_ram_used_gib": 33.98903274536133, "sampled_min_available_phys_gib": 27.666519165039062, "sampled_min_available_commit_gib": 7.800025939941406, "resource_sample_count": 29, "resource_sampling_error_count": 0 }, { "source_study": "iq2-fresh-tail4", "candidate_id": "strata-iq2_xs", "request_id": "performance--speed_graphics_queue_512--2", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 2, "frozen_request_sha256": "5cd334a841fa1a8870459e55b28622255ef481a2257e3a30a274c8a353c65484", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 23.851367800001753, "ttft_any_seconds": 0.7012241999982507, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.7012241999982507, "api_decode_tokens_per_second": 22, "api_prompt_tokens_per_second": 12.7, "sampled_peak_gpu_used_mib": 15660, "sampled_peak_system_ram_used_gib": 33.971221923828125, "sampled_min_available_phys_gib": 27.684329986572266, "sampled_min_available_commit_gib": 7.851642608642578, "resource_sample_count": 23, "resource_sampling_error_count": 0 }, { "source_study": "iq2-fresh-tail4", "candidate_id": "strata-iq2_xs", "request_id": "performance--speed_graphics_queue_512--3", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 3, "frozen_request_sha256": "5cd334a841fa1a8870459e55b28622255ef481a2257e3a30a274c8a353c65484", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 24.462002799999027, "ttft_any_seconds": 0.6500033999982406, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.6500033999982406, "api_decode_tokens_per_second": 21.4, "api_prompt_tokens_per_second": 14.2, "sampled_peak_gpu_used_mib": 15552, "sampled_peak_system_ram_used_gib": 33.98691940307617, "sampled_min_available_phys_gib": 27.66863250732422, "sampled_min_available_commit_gib": 8.016895294189453, "resource_sample_count": 23, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "code--higher_ring_buffer_minimal_fix--0", "case_id": "higher_ring_buffer_minimal_fix", "track": "code", "repeat": 0, "frozen_request_sha256": "1d5cff55c9530ae4506a9831c413c1a15386c52b1c603ea8cd3796f09b3fc4d8", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 278.10923800000455, "ttft_any_seconds": 278.10852370000794, "ttft_reasoning_seconds": 278.10852370000794, "ttft_answer_seconds": 278.10852370000794, "api_decode_tokens_per_second": 11.3, "api_prompt_tokens_per_second": 125.1, "sampled_peak_gpu_used_mib": 15085, "sampled_peak_system_ram_used_gib": 33.95310974121094, "sampled_min_available_phys_gib": 27.702442169189453, "sampled_min_available_commit_gib": 7.452205657958984, "resource_sample_count": 260, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "code--higher_gpu_interval_percentile--0", "case_id": "higher_gpu_interval_percentile", "track": "code", "repeat": 0, "frozen_request_sha256": "af1ff5a4a3e0c8f97ce1f9e2ac7119e51de0f1d58e1f15fbcf288e052c90f508", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 442.7662333000044, "ttft_any_seconds": 442.76416880000033, "ttft_reasoning_seconds": 442.76416880000033, "ttft_answer_seconds": 442.76416880000033, "api_decode_tokens_per_second": 12.3, "api_prompt_tokens_per_second": 122.9, "sampled_peak_gpu_used_mib": 14961, "sampled_peak_system_ram_used_gib": 33.716182708740234, "sampled_min_available_phys_gib": 27.939369201660156, "sampled_min_available_commit_gib": 7.777622222900391, "resource_sample_count": 416, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "code--higher_safe_proposal_validator--0", "case_id": "higher_safe_proposal_validator", "track": "code", "repeat": 0, "frozen_request_sha256": "99f0a9b65b65bc306712174100f9754bb668abf375dddb22da3b67c3f548a8bb", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 420.05647619999945, "ttft_any_seconds": 420.0551242000074, "ttft_reasoning_seconds": 420.0551242000074, "ttft_answer_seconds": 420.0551242000074, "api_decode_tokens_per_second": 14.6, "api_prompt_tokens_per_second": 140.5, "sampled_peak_gpu_used_mib": 14957, "sampled_peak_system_ram_used_gib": 33.28662872314453, "sampled_min_available_phys_gib": 28.36892318725586, "sampled_min_available_commit_gib": 7.8059234619140625, "resource_sample_count": 393, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "code--higher_bilingual_capture_status--0", "case_id": "higher_bilingual_capture_status", "track": "code", "repeat": 0, "frozen_request_sha256": "1b23dfb90c9f89a9d55ef5b135a551215680c6d2c05bb16c0099a7af668993f7", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 449.15604750000057, "ttft_any_seconds": 449.15484520001337, "ttft_reasoning_seconds": 449.15484520001337, "ttft_answer_seconds": 449.15484520001337, "api_decode_tokens_per_second": 11.8, "api_prompt_tokens_per_second": 118.4, "sampled_peak_gpu_used_mib": 14957, "sampled_peak_system_ram_used_gib": 33.197792053222656, "sampled_min_available_phys_gib": 28.457759857177734, "sampled_min_available_commit_gib": 8.074230194091797, "resource_sample_count": 421, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "qa--Q1-Chinese-conversation--0", "case_id": "Q1-Chinese-conversation", "track": "qa", "repeat": 0, "frozen_request_sha256": "55ae05aae2f34bff7c19fb5879a3ea390e04bc8b6aec80136f5c216781f88113", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 43.722214999987045, "ttft_any_seconds": 5.228905700001633, "ttft_reasoning_seconds": 5.228905700001633, "ttft_answer_seconds": 35.586865899997065, "api_decode_tokens_per_second": 10.3, "api_prompt_tokens_per_second": 46.4, "sampled_peak_gpu_used_mib": 14951, "sampled_peak_system_ram_used_gib": 32.77545928955078, "sampled_min_available_phys_gib": 28.88009262084961, "sampled_min_available_commit_gib": 8.575431823730469, "resource_sample_count": 42, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "qa--Q2-Unpaired-percentiles--0", "case_id": "Q2-Unpaired-percentiles", "track": "qa", "repeat": 0, "frozen_request_sha256": "274a5f5e73f4c1f065fea6db1efd81fc89360f96c5958d0472c6ae0de3887b80", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 600.1252396999917, "ttft_any_seconds": 5.241756600007648, "ttft_reasoning_seconds": 5.241756600007648, "ttft_answer_seconds": 354.14261189999525, "api_decode_tokens_per_second": 11.6, "api_prompt_tokens_per_second": 46.1, "sampled_peak_gpu_used_mib": 14952, "sampled_peak_system_ram_used_gib": 32.760555267333984, "sampled_min_available_phys_gib": 28.894996643066406, "sampled_min_available_commit_gib": 8.586006164550781, "resource_sample_count": 562, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "qa--Q3-Overlapping-GPU-scopes--0", "case_id": "Q3-Overlapping-GPU-scopes", "track": "qa", "repeat": 0, "frozen_request_sha256": "a80712aa78419e94afe94ec90f9ac21855a1f68a3a9d9b6b1690a0c40dd23029", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 85.26935079999384, "ttft_any_seconds": 5.5100396999914665, "ttft_reasoning_seconds": 5.5100396999914665, "ttft_answer_seconds": 57.157914300012635, "api_decode_tokens_per_second": 8.7, "api_prompt_tokens_per_second": 58.8, "sampled_peak_gpu_used_mib": 14951, "sampled_peak_system_ram_used_gib": 32.68389892578125, "sampled_min_available_phys_gib": 28.97165298461914, "sampled_min_available_commit_gib": 8.697063446044922, "resource_sample_count": 80, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "qa--Q4-Untrusted-version-advice--0", "case_id": "Q4-Untrusted-version-advice", "track": "qa", "repeat": 0, "frozen_request_sha256": "2619c55b5363b6bab5c8a28d372d6279a7033c743e30f6c97646e95c10dfe3d1", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 581.8775714999938, "ttft_any_seconds": 5.20864209998399, "ttft_reasoning_seconds": 5.20864209998399, "ttft_answer_seconds": 210.18688389999443, "api_decode_tokens_per_second": 12.9, "api_prompt_tokens_per_second": 59.7, "sampled_peak_gpu_used_mib": 14953, "sampled_peak_system_ram_used_gib": 32.99693298339844, "sampled_min_available_phys_gib": 28.658618927001953, "sampled_min_available_commit_gib": 7.903167724609375, "resource_sample_count": 545, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "context--context_lookup_6k_chars--0", "case_id": "context_lookup_6k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "b75ce7ffd5121237f1e2881def920350973139085442b9868e7fb554d44dbd29", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 20.548609700024826, "ttft_any_seconds": 20.547859600017546, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 20.547859600017546, "api_decode_tokens_per_second": 5.6, "api_prompt_tokens_per_second": 436, "sampled_peak_gpu_used_mib": 14946, "sampled_peak_system_ram_used_gib": 32.74326705932617, "sampled_min_available_phys_gib": 28.91228485107422, "sampled_min_available_commit_gib": 8.677642822265625, "resource_sample_count": 20, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "context--context_lookup_24k_chars--0", "case_id": "context_lookup_24k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "2475c0be7488e8f8d409074c6701c211e46bc3042517e76c38969957da7ceb0f", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 25.0976190000074, "ttft_any_seconds": 25.097045800008345, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 25.097045800008345, "api_decode_tokens_per_second": 6.9, "api_prompt_tokens_per_second": 828.9, "sampled_peak_gpu_used_mib": 14946, "sampled_peak_system_ram_used_gib": 32.764671325683594, "sampled_min_available_phys_gib": 28.890880584716797, "sampled_min_available_commit_gib": 8.665409088134766, "resource_sample_count": 24, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "workflow--workflow_no_pie_profile--0", "case_id": "workflow_no_pie_profile", "track": "workflow", "repeat": 0, "frozen_request_sha256": "f570b408a4b612e11073c4dad6686789c28019676057bc9253dc7b9be02cc1cf", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 38.342776499979664, "ttft_any_seconds": 38.3420423999778, "ttft_reasoning_seconds": 38.3420423999778, "ttft_answer_seconds": 38.3420423999778, "api_decode_tokens_per_second": 8.2, "api_prompt_tokens_per_second": 154.5, "sampled_peak_gpu_used_mib": 14946, "sampled_peak_system_ram_used_gib": 32.79558181762695, "sampled_min_available_phys_gib": 28.859970092773438, "sampled_min_available_commit_gib": 8.641731262207031, "resource_sample_count": 36, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "workflow--workflow_untrusted_excerpt--0", "case_id": "workflow_untrusted_excerpt", "track": "workflow", "repeat": 0, "frozen_request_sha256": "7b022df8977e3bff03f1473274a2f64746a8490d14773c25bfae9b471ba7b947", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 123.44726349998382, "ttft_any_seconds": 123.44639239998651, "ttft_reasoning_seconds": 123.44639239998651, "ttft_answer_seconds": 123.44639239998651, "api_decode_tokens_per_second": 10.2, "api_prompt_tokens_per_second": 127.7, "sampled_peak_gpu_used_mib": 14947, "sampled_peak_system_ram_used_gib": 33.01606369018555, "sampled_min_available_phys_gib": 28.639488220214844, "sampled_min_available_commit_gib": 8.324710845947266, "resource_sample_count": 116, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "workflow--workflow_read_only_capture--0", "case_id": "workflow_read_only_capture", "track": "workflow", "repeat": 0, "frozen_request_sha256": "71614835cb2ea22669fc67ad720d4a8db6a1e39ca37d2e997d88b41fe0ee0026", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 64.0597592999984, "ttft_any_seconds": 64.05874909998965, "ttft_reasoning_seconds": 64.05874909998965, "ttft_answer_seconds": 64.05874909998965, "api_decode_tokens_per_second": 8.2, "api_prompt_tokens_per_second": 161.6, "sampled_peak_gpu_used_mib": 14885, "sampled_peak_system_ram_used_gib": 32.89203643798828, "sampled_min_available_phys_gib": 28.76351547241211, "sampled_min_available_commit_gib": 8.6875, "resource_sample_count": 61, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "workflow--workflow_destructive_approval--0", "case_id": "workflow_destructive_approval", "track": "workflow", "repeat": 0, "frozen_request_sha256": "7c03a70c2261bea3ec12b77fb42df47a68e247718421f18ff4eac230e367159e", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 65.08704350001062, "ttft_any_seconds": 65.08621700000367, "ttft_reasoning_seconds": 65.08621700000367, "ttft_answer_seconds": 65.08621700000367, "api_decode_tokens_per_second": 9.7, "api_prompt_tokens_per_second": 156.1, "sampled_peak_gpu_used_mib": 14885, "sampled_peak_system_ram_used_gib": 32.87558364868164, "sampled_min_available_phys_gib": 28.77996826171875, "sampled_min_available_commit_gib": 8.700775146484375, "resource_sample_count": 61, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "performance--speed_graphics_queue_512--1", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 1, "frozen_request_sha256": "9179b478706827e4cd42e21b82b2143bd0e1e3bfc87f00e6e46b92abbe1f4dcd", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 54.16855060000671, "ttft_any_seconds": 4.755726200004574, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 4.755726200004574, "api_decode_tokens_per_second": 10.3, "api_prompt_tokens_per_second": 46, "sampled_peak_gpu_used_mib": 14885, "sampled_peak_system_ram_used_gib": 32.898563385009766, "sampled_min_available_phys_gib": 28.756988525390625, "sampled_min_available_commit_gib": 8.666709899902344, "resource_sample_count": 51, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "performance--speed_graphics_queue_512--2", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 2, "frozen_request_sha256": "9179b478706827e4cd42e21b82b2143bd0e1e3bfc87f00e6e46b92abbe1f4dcd", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 40.43938310001977, "ttft_any_seconds": 0.8293482000008225, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.8293482000008225, "api_decode_tokens_per_second": 12.9, "api_prompt_tokens_per_second": 11.9, "sampled_peak_gpu_used_mib": 14885, "sampled_peak_system_ram_used_gib": 32.863922119140625, "sampled_min_available_phys_gib": 28.791629791259766, "sampled_min_available_commit_gib": 8.722259521484375, "resource_sample_count": 38, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_xxs", "request_id": "performance--speed_graphics_queue_512--3", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 3, "frozen_request_sha256": "9179b478706827e4cd42e21b82b2143bd0e1e3bfc87f00e6e46b92abbe1f4dcd", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 40.82330469999579, "ttft_any_seconds": 0.8674681000120472, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.8674681000120472, "api_decode_tokens_per_second": 12.7, "api_prompt_tokens_per_second": 11.3, "sampled_peak_gpu_used_mib": 14885, "sampled_peak_system_ram_used_gib": 32.89594650268555, "sampled_min_available_phys_gib": 28.759605407714844, "sampled_min_available_commit_gib": 8.688179016113281, "resource_sample_count": 39, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "code--higher_ring_buffer_minimal_fix--0", "case_id": "higher_ring_buffer_minimal_fix", "track": "code", "repeat": 0, "frozen_request_sha256": "56225f5cf9502daa8f5c90a75340bb508c3562bd68624149adf8de20d2c293b6", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 184.699059400009, "ttft_any_seconds": 184.69823050001287, "ttft_reasoning_seconds": 184.69823050001287, "ttft_answer_seconds": 184.69823050001287, "api_decode_tokens_per_second": 8.3, "api_prompt_tokens_per_second": 112.2, "sampled_peak_gpu_used_mib": 14965, "sampled_peak_system_ram_used_gib": 33.64236068725586, "sampled_min_available_phys_gib": 28.01319122314453, "sampled_min_available_commit_gib": 7.812480926513672, "resource_sample_count": 173, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "code--higher_gpu_interval_percentile--0", "case_id": "higher_gpu_interval_percentile", "track": "code", "repeat": 0, "frozen_request_sha256": "ad503c632147d06c33e36728dfed935ca303fffe863e71155c1edf52b6a53735", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 585.4577757000225, "ttft_any_seconds": 585.4566042000079, "ttft_reasoning_seconds": 585.4566042000079, "ttft_answer_seconds": 585.4566042000079, "api_decode_tokens_per_second": 9.4, "api_prompt_tokens_per_second": 62.6, "sampled_peak_gpu_used_mib": 14965, "sampled_peak_system_ram_used_gib": 33.73042297363281, "sampled_min_available_phys_gib": 27.925128936767578, "sampled_min_available_commit_gib": 7.755275726318359, "resource_sample_count": 547, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "code--higher_safe_proposal_validator--0", "case_id": "higher_safe_proposal_validator", "track": "code", "repeat": 0, "frozen_request_sha256": "fbd64f965a6e223303af0849031f1366c78f380a77d98b8b835c7303368a6400", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 576.2479724999866, "ttft_any_seconds": 576.2470924000081, "ttft_reasoning_seconds": 576.2470924000081, "ttft_answer_seconds": 576.2470924000081, "api_decode_tokens_per_second": 10.3, "api_prompt_tokens_per_second": 124.4, "sampled_peak_gpu_used_mib": 14963, "sampled_peak_system_ram_used_gib": 34.27897262573242, "sampled_min_available_phys_gib": 27.37657928466797, "sampled_min_available_commit_gib": 7.176532745361328, "resource_sample_count": 538, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "code--higher_bilingual_capture_status--0", "case_id": "higher_bilingual_capture_status", "track": "code", "repeat": 0, "frozen_request_sha256": "0d644d29a60f6e331f8b57493b1d018c47ef6857afd8a39574373926f2581a4b", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 590.3191635000112, "ttft_any_seconds": 590.3178200000257, "ttft_reasoning_seconds": 590.3178200000257, "ttft_answer_seconds": 590.3178200000257, "api_decode_tokens_per_second": 8.9, "api_prompt_tokens_per_second": 105, "sampled_peak_gpu_used_mib": 15036, "sampled_peak_system_ram_used_gib": 34.145545959472656, "sampled_min_available_phys_gib": 27.510005950927734, "sampled_min_available_commit_gib": 7.334377288818359, "resource_sample_count": 551, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "qa--Q1-Chinese-conversation--0", "case_id": "Q1-Chinese-conversation", "track": "qa", "repeat": 0, "frozen_request_sha256": "0caef9d7d2ab437ecafb7a16369c927beaaecc57ac7a9f53945c8ca9694e1558", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 43.17396580000059, "ttft_any_seconds": 6.061072400014382, "ttft_reasoning_seconds": 6.061072400014382, "ttft_answer_seconds": 24.666014200018253, "api_decode_tokens_per_second": 7.4, "api_prompt_tokens_per_second": 40.3, "sampled_peak_gpu_used_mib": 15015, "sampled_peak_system_ram_used_gib": 33.93546676635742, "sampled_min_available_phys_gib": 27.72008514404297, "sampled_min_available_commit_gib": 7.510669708251953, "resource_sample_count": 41, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "qa--Q2-Unpaired-percentiles--0", "case_id": "Q2-Unpaired-percentiles", "track": "qa", "repeat": 0, "frozen_request_sha256": "418836046f9c3f72fc97f6f5a386ce4cb078cfb549cb755fefd7cfb389ad9a17", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 269.84436869999627, "ttft_any_seconds": 6.05128159999731, "ttft_reasoning_seconds": 6.05128159999731, "ttft_answer_seconds": 146.05102610000176, "api_decode_tokens_per_second": 8.5, "api_prompt_tokens_per_second": 50, "sampled_peak_gpu_used_mib": 15579, "sampled_peak_system_ram_used_gib": 35.2878303527832, "sampled_min_available_phys_gib": 26.367721557617188, "sampled_min_available_commit_gib": 6.01025390625, "resource_sample_count": 253, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "qa--Q3-Overlapping-GPU-scopes--0", "case_id": "Q3-Overlapping-GPU-scopes", "track": "qa", "repeat": 0, "frozen_request_sha256": "95400961101fc69b9e24f839428376cc2837f31c49b1199cf749c7decc5d8c66", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 94.48075730001437, "ttft_any_seconds": 6.361200300016208, "ttft_reasoning_seconds": 6.361200300016208, "ttft_answer_seconds": 65.93494549999014, "api_decode_tokens_per_second": 6.8, "api_prompt_tokens_per_second": 51.3, "sampled_peak_gpu_used_mib": 15209, "sampled_peak_system_ram_used_gib": 35.25656509399414, "sampled_min_available_phys_gib": 26.39898681640625, "sampled_min_available_commit_gib": 6.155250549316406, "resource_sample_count": 89, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "qa--Q4-Untrusted-version-advice--0", "case_id": "Q4-Untrusted-version-advice", "track": "qa", "repeat": 0, "frozen_request_sha256": "51ee27eed5e17676a04c50fae5ce96d961b455b2e380a32218be61673a42077a", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 686.0040467000217, "ttft_any_seconds": 6.070143800025107, "ttft_reasoning_seconds": 6.070143800025107, "ttft_answer_seconds": 322.16883640000015, "api_decode_tokens_per_second": 9.4, "api_prompt_tokens_per_second": 51.2, "sampled_peak_gpu_used_mib": 15477, "sampled_peak_system_ram_used_gib": 35.76994323730469, "sampled_min_available_phys_gib": 25.885608673095703, "sampled_min_available_commit_gib": 4.744968414306641, "resource_sample_count": 636, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "context--context_lookup_6k_chars--0", "case_id": "context_lookup_6k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "99df96c14c6b0baf6c62a622f666ecb7701fd59d23923494f71ce88651c43d4f", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 23.97752089999267, "ttft_any_seconds": 23.97694219998084, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 23.97694219998084, "api_decode_tokens_per_second": 4.7, "api_prompt_tokens_per_second": 384.2, "sampled_peak_gpu_used_mib": 15085, "sampled_peak_system_ram_used_gib": 34.61995315551758, "sampled_min_available_phys_gib": 27.035598754882812, "sampled_min_available_commit_gib": 6.454555511474609, "resource_sample_count": 23, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "context--context_lookup_24k_chars--0", "case_id": "context_lookup_24k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "94979b802069a26c073af44664fc88c16f28d3ea78318e78231415197ed7708d", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 28.5204192000092, "ttft_any_seconds": 28.519758800015552, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 28.519758800015552, "api_decode_tokens_per_second": 6.1, "api_prompt_tokens_per_second": 731.5, "sampled_peak_gpu_used_mib": 15087, "sampled_peak_system_ram_used_gib": 34.733489990234375, "sampled_min_available_phys_gib": 26.922061920166016, "sampled_min_available_commit_gib": 6.251060485839844, "resource_sample_count": 27, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "workflow--workflow_no_pie_profile--0", "case_id": "workflow_no_pie_profile", "track": "workflow", "repeat": 0, "frozen_request_sha256": "213784369ef27ed8ce34e7fe5f4730248486a10e929cc69d29da8db3ec7acf90", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 51.765128199971514, "ttft_any_seconds": 51.7644764999859, "ttft_reasoning_seconds": 51.7644764999859, "ttft_answer_seconds": 51.7644764999859, "api_decode_tokens_per_second": 6.2, "api_prompt_tokens_per_second": 132.2, "sampled_peak_gpu_used_mib": 15159, "sampled_peak_system_ram_used_gib": 34.770565032958984, "sampled_min_available_phys_gib": 26.884986877441406, "sampled_min_available_commit_gib": 6.190544128417969, "resource_sample_count": 49, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "workflow--workflow_untrusted_excerpt--0", "case_id": "workflow_untrusted_excerpt", "track": "workflow", "repeat": 0, "frozen_request_sha256": "917dfd9eacee19cc2bd4c52cced08c89a30857a121cc48364437c7acc29218c5", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 153.62164550001035, "ttft_any_seconds": 153.62068950000685, "ttft_reasoning_seconds": 153.62068950000685, "ttft_answer_seconds": 153.62068950000685, "api_decode_tokens_per_second": 7.3, "api_prompt_tokens_per_second": 136, "sampled_peak_gpu_used_mib": 15195, "sampled_peak_system_ram_used_gib": 35.23013687133789, "sampled_min_available_phys_gib": 26.4254150390625, "sampled_min_available_commit_gib": 5.840854644775391, "resource_sample_count": 144, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "workflow--workflow_read_only_capture--0", "case_id": "workflow_read_only_capture", "track": "workflow", "repeat": 0, "frozen_request_sha256": "44f00eb12a5334f2e0fa8a4680a7350003e780028fd9fc114060d33198e88798", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 116.54225190001307, "ttft_any_seconds": 116.54144160001306, "ttft_reasoning_seconds": 116.54144160001306, "ttft_answer_seconds": 116.54144160001306, "api_decode_tokens_per_second": 6.8, "api_prompt_tokens_per_second": 135.5, "sampled_peak_gpu_used_mib": 15131, "sampled_peak_system_ram_used_gib": 35.07122039794922, "sampled_min_available_phys_gib": 26.584331512451172, "sampled_min_available_commit_gib": 5.9672698974609375, "resource_sample_count": 109, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "workflow--workflow_destructive_approval--0", "case_id": "workflow_destructive_approval", "track": "workflow", "repeat": 0, "frozen_request_sha256": "233199aef8f24bd9e34ef67acbd2cb798acf5b173366e7aadc2dee11532b4edc", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 108.98173559998395, "ttft_any_seconds": 108.98098520000349, "ttft_reasoning_seconds": 108.98098520000349, "ttft_answer_seconds": 108.98098520000349, "api_decode_tokens_per_second": 7.2, "api_prompt_tokens_per_second": 135.6, "sampled_peak_gpu_used_mib": 15150, "sampled_peak_system_ram_used_gib": 35.328792572021484, "sampled_min_available_phys_gib": 26.326759338378906, "sampled_min_available_commit_gib": 5.720161437988281, "resource_sample_count": 102, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "performance--speed_graphics_queue_512--1", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 1, "frozen_request_sha256": "e7a2be0da1df9b97bab1755e719726731a18cda2142b6c9b04eb85b82f013371", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 67.45779540002695, "ttft_any_seconds": 5.720312800025567, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 5.720312800025567, "api_decode_tokens_per_second": 8.3, "api_prompt_tokens_per_second": 38.5, "sampled_peak_gpu_used_mib": 15131, "sampled_peak_system_ram_used_gib": 35.319759368896484, "sampled_min_available_phys_gib": 26.335792541503906, "sampled_min_available_commit_gib": 5.733436584472656, "resource_sample_count": 64, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "performance--speed_graphics_queue_512--2", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 2, "frozen_request_sha256": "e7a2be0da1df9b97bab1755e719726731a18cda2142b6c9b04eb85b82f013371", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 64.28774090000661, "ttft_any_seconds": 0.9798636000195984, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.9798636000195984, "api_decode_tokens_per_second": 8.1, "api_prompt_tokens_per_second": 9.5, "sampled_peak_gpu_used_mib": 15064, "sampled_peak_system_ram_used_gib": 35.163230895996094, "sampled_min_available_phys_gib": 26.492321014404297, "sampled_min_available_commit_gib": 5.943611145019531, "resource_sample_count": 60, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-iq3_s", "request_id": "performance--speed_graphics_queue_512--3", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 3, "frozen_request_sha256": "e7a2be0da1df9b97bab1755e719726731a18cda2142b6c9b04eb85b82f013371", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 59.470488499995554, "ttft_any_seconds": 0.9598735999898054, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 0.9598735999898054, "api_decode_tokens_per_second": 8.7, "api_prompt_tokens_per_second": 10, "sampled_peak_gpu_used_mib": 15058, "sampled_peak_system_ram_used_gib": 35.07213592529297, "sampled_min_available_phys_gib": 26.583415985107422, "sampled_min_available_commit_gib": 6.032478332519531, "resource_sample_count": 55, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-ud-q4_k_xl", "request_id": "code--higher_ring_buffer_minimal_fix--0", "case_id": "higher_ring_buffer_minimal_fix", "track": "code", "repeat": 0, "frozen_request_sha256": "5292136914954040f4aab76a46efde8caeb5a8a37ec5d0016f6b7c0778091bb3", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 524.4812079999829, "ttft_any_seconds": 524.4451346999849, "ttft_reasoning_seconds": 524.4451346999849, "ttft_answer_seconds": 524.4451346999849, "api_decode_tokens_per_second": 3.2, "api_prompt_tokens_per_second": 62.7, "sampled_peak_gpu_used_mib": 15322, "sampled_peak_system_ram_used_gib": 33.40501403808594, "sampled_min_available_phys_gib": 28.250537872314453, "sampled_min_available_commit_gib": 7.7109527587890625, "resource_sample_count": 489, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-ud-q4_k_xl", "request_id": "code--higher_gpu_interval_percentile--0", "case_id": "higher_gpu_interval_percentile", "track": "code", "repeat": 0, "frozen_request_sha256": "3422c7696d2c4772a8e1c052992936e2c1ab238f9bee6f4ed7a23661ae1539d9", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 1562.3865867000131, "ttft_any_seconds": 1562.3851740999962, "ttft_reasoning_seconds": 1562.3851740999962, "ttft_answer_seconds": 1562.3851740999962, "api_decode_tokens_per_second": 3.3, "api_prompt_tokens_per_second": 58.5, "sampled_peak_gpu_used_mib": 15212, "sampled_peak_system_ram_used_gib": 31.459190368652344, "sampled_min_available_phys_gib": 30.196361541748047, "sampled_min_available_commit_gib": 8.074565887451172, "resource_sample_count": 1452, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-ud-q4_k_xl", "request_id": "code--higher_safe_proposal_validator--0", "case_id": "higher_safe_proposal_validator", "track": "code", "repeat": 0, "frozen_request_sha256": "4e83c7c544a996d32282dde9c5fa17c1666a820b3b88f4d210c05995ae00f510", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 1657.669366200018, "ttft_any_seconds": 1657.668345300015, "ttft_reasoning_seconds": 1657.668345300015, "ttft_answer_seconds": 1657.668345300015, "api_decode_tokens_per_second": 3.6, "api_prompt_tokens_per_second": 68.8, "sampled_peak_gpu_used_mib": 15165, "sampled_peak_system_ram_used_gib": 31.421844482421875, "sampled_min_available_phys_gib": 30.233707427978516, "sampled_min_available_commit_gib": 8.21847915649414, "resource_sample_count": 1553, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-ud-q4_k_xl", "request_id": "code--higher_bilingual_capture_status--0", "case_id": "higher_bilingual_capture_status", "track": "code", "repeat": 0, "frozen_request_sha256": "11eb46f38dc231a6c21082a2648bbe7ff93cffa2f068daec72773516cfec4a1f", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 1569.4152174000046, "ttft_any_seconds": 1569.4137749999936, "ttft_reasoning_seconds": 1569.4137749999936, "ttft_answer_seconds": 1569.4137749999936, "api_decode_tokens_per_second": 3.4, "api_prompt_tokens_per_second": 41.5, "sampled_peak_gpu_used_mib": 15699, "sampled_peak_system_ram_used_gib": 31.990909576416016, "sampled_min_available_phys_gib": 29.664642333984375, "sampled_min_available_commit_gib": 6.653133392333984, "resource_sample_count": 1464, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-ud-q4_k_xl", "request_id": "qa--Q1-Chinese-conversation--0", "case_id": "Q1-Chinese-conversation", "track": "qa", "repeat": 0, "frozen_request_sha256": "65d321e01b43d50b62cc8a504fc018d6327e13387b0abdabdac122d1861ba6e3", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 92.18451089999871, "ttft_any_seconds": 10.127458800008753, "ttft_reasoning_seconds": 10.127458800008753, "ttft_answer_seconds": 48.280332800000906, "api_decode_tokens_per_second": 3, "api_prompt_tokens_per_second": 24.4, "sampled_peak_gpu_used_mib": 15272, "sampled_peak_system_ram_used_gib": 31.278282165527344, "sampled_min_available_phys_gib": 30.377269744873047, "sampled_min_available_commit_gib": 7.172405242919922, "resource_sample_count": 86, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-ud-q4_k_xl", "request_id": "qa--Q2-Unpaired-percentiles--0", "case_id": "Q2-Unpaired-percentiles", "track": "qa", "repeat": 0, "frozen_request_sha256": "aea051302731eb16067e85de754315d2775c0e926481cf2f1ea023f5cd6266c6", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 1037.8838942000002, "ttft_any_seconds": 10.32419139999547, "ttft_reasoning_seconds": 10.32419139999547, "ttft_answer_seconds": 481.539432199992, "api_decode_tokens_per_second": 3.2, "api_prompt_tokens_per_second": 29.5, "sampled_peak_gpu_used_mib": 15127, "sampled_peak_system_ram_used_gib": 31.388996124267578, "sampled_min_available_phys_gib": 30.266555786132812, "sampled_min_available_commit_gib": 7.509307861328125, "resource_sample_count": 968, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-ud-q4_k_xl", "request_id": "qa--Q3-Overlapping-GPU-scopes--0", "case_id": "Q3-Overlapping-GPU-scopes", "track": "qa", "repeat": 0, "frozen_request_sha256": "4a8d00491673a62ae01b6e2503a321b8787c700e098c1ab819a2d632e888ccf4", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": true, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 246.80575739999767, "ttft_any_seconds": 10.969806499982951, "ttft_reasoning_seconds": 10.969806499982951, "ttft_answer_seconds": 191.18625249998877, "api_decode_tokens_per_second": 3, "api_prompt_tokens_per_second": 29.8, "sampled_peak_gpu_used_mib": 15103, "sampled_peak_system_ram_used_gib": 31.369678497314453, "sampled_min_available_phys_gib": 30.285873413085938, "sampled_min_available_commit_gib": 6.931591033935547, "resource_sample_count": 231, "resource_sampling_error_count": 0 }, { "source_study": "original-v3", "candidate_id": "strata-ud-q4_k_xl", "request_id": "qa--Q4-Untrusted-version-advice--0", "case_id": "Q4-Untrusted-version-advice", "track": "qa", "repeat": 0, "frozen_request_sha256": "33789fbaf8582ff735922a610d9a69fd08779a8245eb5c17c9af226c0a000e73", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "transport_or_response_failure", "formal_metrics_eligible": false, "diagnostic_only": true, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": null, "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 12288, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 4096, "wall_seconds": 1800.5894061000145, "ttft_any_seconds": 10.267801000009058, "ttft_reasoning_seconds": 10.267801000009058, "ttft_answer_seconds": 987.6406314000196, "api_decode_tokens_per_second": null, "api_prompt_tokens_per_second": null, "sampled_peak_gpu_used_mib": 15192, "sampled_peak_system_ram_used_gib": 31.460086822509766, "sampled_min_available_phys_gib": 30.195465087890625, "sampled_min_available_commit_gib": 7.326595306396484, "resource_sample_count": 1666, "resource_sampling_error_count": 0 }, { "source_study": "q4-tail9", "candidate_id": "strata-ud-q4_k_xl", "request_id": "context--context_lookup_6k_chars--0", "case_id": "context_lookup_6k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "69e183571610f8a8c7ea5a86b9785454982fcf31ec8fed12e271ccaff8ae9769", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 38.16839869999967, "ttft_any_seconds": 38.16778730000078, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 38.16778730000078, "api_decode_tokens_per_second": 3.3, "api_prompt_tokens_per_second": 209.9, "sampled_peak_gpu_used_mib": 15691, "sampled_peak_system_ram_used_gib": 39.836578369140625, "sampled_min_available_phys_gib": 21.818973541259766, "sampled_min_available_commit_gib": 6.85345458984375, "resource_sample_count": 36, "resource_sampling_error_count": 0 }, { "source_study": "q4-tail9", "candidate_id": "strata-ud-q4_k_xl", "request_id": "context--context_lookup_24k_chars--0", "case_id": "context_lookup_24k_chars", "track": "context", "repeat": 0, "frozen_request_sha256": "0ae80d56b7eb87563db83253e9242d0a993c73449793fba419cf77e7fda7e01e", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 45.731234400001995, "ttft_any_seconds": 45.73067450000235, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 45.73067450000235, "api_decode_tokens_per_second": 4.2, "api_prompt_tokens_per_second": 433.1, "sampled_peak_gpu_used_mib": 15641, "sampled_peak_system_ram_used_gib": 39.525306701660156, "sampled_min_available_phys_gib": 22.130245208740234, "sampled_min_available_commit_gib": 7.172344207763672, "resource_sample_count": 43, "resource_sampling_error_count": 0 }, { "source_study": "q4-tail9", "candidate_id": "strata-ud-q4_k_xl", "request_id": "workflow--workflow_no_pie_profile--0", "case_id": "workflow_no_pie_profile", "track": "workflow", "repeat": 0, "frozen_request_sha256": "d9a5c1a2360fb7fc56e2bd0513e54a52271c2deb47c5fbb33b0299716b071e20", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 183.6751321000047, "ttft_any_seconds": 183.6741403000051, "ttft_reasoning_seconds": 183.6741403000051, "ttft_answer_seconds": 183.6741403000051, "api_decode_tokens_per_second": 4.1, "api_prompt_tokens_per_second": 79.8, "sampled_peak_gpu_used_mib": 15658, "sampled_peak_system_ram_used_gib": 36.5853271484375, "sampled_min_available_phys_gib": 25.07022476196289, "sampled_min_available_commit_gib": 6.371257781982422, "resource_sample_count": 170, "resource_sampling_error_count": 0 }, { "source_study": "q4-tail9", "candidate_id": "strata-ud-q4_k_xl", "request_id": "workflow--workflow_untrusted_excerpt--0", "case_id": "workflow_untrusted_excerpt", "track": "workflow", "repeat": 0, "frozen_request_sha256": "b732909772dd30a390b3ea98e3d48e4e6e2103988d0abf5f61da132035e3409c", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 262.77222690000053, "ttft_any_seconds": 262.77147310000146, "ttft_reasoning_seconds": 262.77147310000146, "ttft_answer_seconds": 262.77147310000146, "api_decode_tokens_per_second": 3.8, "api_prompt_tokens_per_second": 86, "sampled_peak_gpu_used_mib": 15720, "sampled_peak_system_ram_used_gib": 39.513282775878906, "sampled_min_available_phys_gib": 22.142269134521484, "sampled_min_available_commit_gib": 2.4899673461914062, "resource_sample_count": 244, "resource_sampling_error_count": 0 }, { "source_study": "q4-tail9", "candidate_id": "strata-ud-q4_k_xl", "request_id": "workflow--workflow_read_only_capture--0", "case_id": "workflow_read_only_capture", "track": "workflow", "repeat": 0, "frozen_request_sha256": "5304ef29506b99a80817fcd152bdc9490c0161b7eb6b51c1a886ca942e55fbf6", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": true, "diagnostic_deterministic_fixture_passed": true, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 236.35692439999548, "ttft_any_seconds": 236.35604509999393, "ttft_reasoning_seconds": 236.35604509999393, "ttft_answer_seconds": 236.35604509999393, "api_decode_tokens_per_second": 3.9, "api_prompt_tokens_per_second": 88.6, "sampled_peak_gpu_used_mib": 15647, "sampled_peak_system_ram_used_gib": 38.570350646972656, "sampled_min_available_phys_gib": 23.085201263427734, "sampled_min_available_commit_gib": 4.002296447753906, "resource_sample_count": 220, "resource_sampling_error_count": 0 }, { "source_study": "q4-tail9", "candidate_id": "strata-ud-q4_k_xl", "request_id": "workflow--workflow_destructive_approval--0", "case_id": "workflow_destructive_approval", "track": "workflow", "repeat": 0, "frozen_request_sha256": "679f36547574abeaba7b9e3ae0368cbd9c602f91a147cb2176d634d38679a212", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "completed", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": true, "observed_complete_response_delivery": true, "schema_validation": "passed", "formal_deterministic_fixture_passed": false, "diagnostic_deterministic_fixture_passed": false, "deterministic_check_observed": true, "max_generated_tokens_requested": 4096, "reasoning_effort_requested": "high", "reasoning_budget_tokens_requested": 1024, "wall_seconds": 167.7044001000031, "ttft_any_seconds": 167.70367790000455, "ttft_reasoning_seconds": 167.70367790000455, "ttft_answer_seconds": 167.70367790000455, "api_decode_tokens_per_second": 3.8, "api_prompt_tokens_per_second": 87.4, "sampled_peak_gpu_used_mib": 15541, "sampled_peak_system_ram_used_gib": 38.32509231567383, "sampled_min_available_phys_gib": 23.330459594726562, "sampled_min_available_commit_gib": 3.6919288635253906, "resource_sample_count": 156, "resource_sampling_error_count": 0 }, { "source_study": "q4-tail9", "candidate_id": "strata-ud-q4_k_xl", "request_id": "performance--speed_graphics_queue_512--1", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 1, "frozen_request_sha256": "01fed2ca771641cbf43cf5f20061088176a53eec442cf9a9dd82d0074d27b207", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 134.40233509999962, "ttft_any_seconds": 8.807977199998277, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 8.807977199998277, "api_decode_tokens_per_second": 4.1, "api_prompt_tokens_per_second": 25.2, "sampled_peak_gpu_used_mib": 15510, "sampled_peak_system_ram_used_gib": 38.98678970336914, "sampled_min_available_phys_gib": 22.66876220703125, "sampled_min_available_commit_gib": 3.28875732421875, "resource_sample_count": 124, "resource_sampling_error_count": 0 }, { "source_study": "q4-tail9", "candidate_id": "strata-ud-q4_k_xl", "request_id": "performance--speed_graphics_queue_512--2", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 2, "frozen_request_sha256": "01fed2ca771641cbf43cf5f20061088176a53eec442cf9a9dd82d0074d27b207", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 132.54401130000042, "ttft_any_seconds": 1.6631378999954904, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.6631378999954904, "api_decode_tokens_per_second": 3.9, "api_prompt_tokens_per_second": 5.5, "sampled_peak_gpu_used_mib": 15580, "sampled_peak_system_ram_used_gib": 38.27993392944336, "sampled_min_available_phys_gib": 23.37561798095703, "sampled_min_available_commit_gib": 3.787914276123047, "resource_sample_count": 124, "resource_sampling_error_count": 0 }, { "source_study": "q4-tail9", "candidate_id": "strata-ud-q4_k_xl", "request_id": "performance--speed_graphics_queue_512--3", "case_id": "speed_graphics_queue_512", "track": "performance", "repeat": 3, "frozen_request_sha256": "01fed2ca771641cbf43cf5f20061088176a53eec442cf9a9dd82d0074d27b207", "attempted_or_intent_recorded": true, "response_recorded": true, "pending_unfinalized": false, "response_status": "truncated", "formal_metrics_eligible": true, "diagnostic_only": false, "formal_complete_response_delivery": false, "observed_complete_response_delivery": false, "schema_validation": "not_requested", "formal_deterministic_fixture_passed": null, "diagnostic_deterministic_fixture_passed": null, "deterministic_check_observed": false, "max_generated_tokens_requested": 512, "reasoning_effort_requested": "none", "reasoning_budget_tokens_requested": 0, "wall_seconds": 136.0531692999939, "ttft_any_seconds": 1.6623418999952264, "ttft_reasoning_seconds": null, "ttft_answer_seconds": 1.6623418999952264, "api_decode_tokens_per_second": 3.8, "api_prompt_tokens_per_second": 5.6, "sampled_peak_gpu_used_mib": 15470, "sampled_peak_system_ram_used_gib": 38.131431579589844, "sampled_min_available_phys_gib": 23.524120330810547, "sampled_min_available_commit_gib": 4.096549987792969, "resource_sample_count": 127, "resource_sampling_error_count": 0 } ], "runtime_version_evidence": [ { "source_study": "original-v3", "candidate_id": "ollama-gptoss-ggml", "measured_owned_server_version": "0.40.2", "actual_runner": "llamacpp", "recorded_manifest_label": "ggml", "source_artifact": "ollama-gptoss-ggml/server.stderr.log", "source_log_sha256": "72fd83a97a4eb1306f39ce3e493deeb01b3517548898bbd047db3b91e1c8ae19", "recorded_source_log_sha256": "72fd83a97a4eb1306f39ce3e493deeb01b3517548898bbd047db3b91e1c8ae19", "source_log_bytes": 130692, "recorded_digest_matches": true }, { "source_study": "original-v3", "candidate_id": "ollama-gptoss-llamacpp", "measured_owned_server_version": "0.40.2", "actual_runner": "llamacpp", "recorded_manifest_label": "llamacpp", "source_artifact": "ollama-gptoss-llamacpp/server.stderr.log", "source_log_sha256": "e7279d0d13ea21e7f19bb2de6789a5e715c1984943043d650a822c977598e770", "recorded_source_log_sha256": "e7279d0d13ea21e7f19bb2de6789a5e715c1984943043d650a822c977598e770", "source_log_bytes": 128579, "recorded_digest_matches": true }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-ggml", "measured_owned_server_version": "0.40.2", "actual_runner": "llamacpp", "recorded_manifest_label": "ggml", "source_artifact": "ollama-qwen35-ggml/server.stderr.log", "source_log_sha256": "4de4e90fb646c676ddaff918373e178454df8608708242829d769cb67c612bdb", "recorded_source_log_sha256": "4de4e90fb646c676ddaff918373e178454df8608708242829d769cb67c612bdb", "source_log_bytes": 176998, "recorded_digest_matches": true }, { "source_study": "original-v3", "candidate_id": "ollama-qwen35-llamacpp", "measured_owned_server_version": "0.40.2", "actual_runner": "llamacpp", "recorded_manifest_label": "llamacpp", "source_artifact": "ollama-qwen35-llamacpp/server.stderr.log", "source_log_sha256": "7c76231617331f8358d434ec727cc958afff58973b51ba49cab9a09f9e4ac32f", "recorded_source_log_sha256": "7c76231617331f8358d434ec727cc958afff58973b51ba49cab9a09f9e4ac32f", "source_log_bytes": 165856, "recorded_digest_matches": true }, { "source_study": "original-v3", "candidate_id": "ollama-qwen3vl", "measured_owned_server_version": "0.40.2", "actual_runner": "llamacpp", "recorded_manifest_label": "ggml", "source_artifact": "ollama-qwen3vl/server.stderr.log", "source_log_sha256": "d4acb0379f5dcbb35ede42e347a6b85bb7e653d0d9418784d14fbd4f79e47760", "recorded_source_log_sha256": "d4acb0379f5dcbb35ede42e347a6b85bb7e653d0d9418784d14fbd4f79e47760", "source_log_bytes": 137973, "recorded_digest_matches": true }, { "source_study": "original-v3", "candidate_id": "ollama-embedding", "measured_owned_server_version": "0.40.2", "actual_runner": "llamacpp", "recorded_manifest_label": "ggml", "source_artifact": "ollama-embedding/server.stderr.log", "source_log_sha256": "ad4da58b8345d46687c0fc846952bb130de54f8d523d61fb7f147d6723889ac8", "recorded_source_log_sha256": "ad4da58b8345d46687c0fc846952bb130de54f8d523d61fb7f147d6723889ac8", "source_log_bytes": 40065, "recorded_digest_matches": true } ], "metric_definitions": { "response_record": "A persisted response measurement, including failed or truncated records; not necessarily a complete answer.", "formal_complete_response_delivery": "Complete response delivery in a separately completed, lifecycle-qualified sealed segment. Delivery does not imply schema/fixture correctness or human acceptance.", "diagnostic_complete_response_delivery_observed": "Observed complete delivery, including unqualified diagnostics; not a formal quality sample.", "whole_case_code_pass": "All required bounded checks and the frozen interface constraints pass for one complete coding case.", "executed_code_check": "A bounded assertion actually reached execution. Unexecuted checks remain visible separately and are not counted as executed failures.", "agent_rubric_acceptance": "Preliminary blind agent rubric decision on authored QA fixtures; not human acceptance or production acceptance.", "decode_tokens_per_second": "API-reported output decode throughput for the configured system and frozen sample. Text speed is a capped 512-token sample, not complete-answer latency.", "ttft_any_seconds": "First streaming protocol chunk; may differ from first visible answer.", "ttft_answer_seconds": "First visible answer content, distinct from protocol events and reasoning. Structured buffering/template differences limit direct interpretation.", "residency_cold_load": "Model residency cold only; not disk or operating-system cache cold.", "whole_system_lifecycle_samples": "About 1 Hz whole-system RAM, VRAM and system-commit-headroom observations during the specified lifecycle. Not model-exclusive allocation or exact transient peaks.", "unavailable_metric": "null or a non-planned track means not measured/not planned, not a zero-quality score.", "study_isolation": "Never pool old failures, interrupted records or independent tails into a single candidate-level performance or quality score." }, "publication_date": "2026-10-12" }