{ "generated_at": "2026-09-07T01:40:05.321416+00:00", "run_count": 30, "runs": [ { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": {}, "display_name": "cydonia-24b-v4.3", "efficiency": { "cold_start_seconds": 27.967421191999165, "cold_ttft_seconds": 0.005088760997750796, "median_task_seconds": 37.183, "tasks_per_minute": 1.6136, "total_seconds": 37.183, "workload_seconds": 9.215588352999475, "workload_ttft_seconds": 2.350872294999135 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 0, "core": 1, "strict": 0, "total": 1 }, "failure_counts": { "tool_protocol_failure": 1 }, "manifest_path": "runs/20260812t015120-latency-cydonia-24b-v4-3-r1/manifest.json", "model": "cydonia-24b-v4.3", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 1, "evaluated_pass_rate": 0.0, "scheduled": 1, "scheduled_pass_rate": 0.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260812t015120-latency-cydonia-24b-v4-3-r1", "score": 50.0, "started_at": "2026-08-12T01:51:20.432340-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260812t015120-latency-cydonia-24b-v4-3-r1/summary.json", "task_results_path": "runs/20260812t015120-latency-cydonia-24b-v4-3-r1/task-results.jsonl" }, { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": {}, "display_name": "dolphin-mistral-24b-venice", "efficiency": { "cold_start_seconds": 17.65202158899774, "cold_ttft_seconds": 0.0032104339989018627, "median_task_seconds": 18.6726, "tasks_per_minute": 3.2133, "total_seconds": 18.6726, "workload_seconds": 1.020600588999514, "workload_ttft_seconds": 0.6275144590035779 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 0, "core": 1, "strict": 0, "total": 1 }, "failure_counts": { "tool_protocol_failure": 1 }, "manifest_path": "runs/20260812t015120-latency-dolphin-mistral-24b-venice-r1/manifest.json", "model": "dolphin-mistral-24b-venice", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 1, "evaluated_pass_rate": 0.0, "scheduled": 1, "scheduled_pass_rate": 0.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260812t015120-latency-dolphin-mistral-24b-venice-r1", "score": 50.0, "started_at": "2026-08-12T01:51:20.432340-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260812t015120-latency-dolphin-mistral-24b-venice-r1/summary.json", "task_results_path": "runs/20260812t015120-latency-dolphin-mistral-24b-venice-r1/task-results.jsonl" }, { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": {}, "display_name": "empero-ai-qwythos-9b-claude-mythos-5-1m-gguf-qwythos-9b-claude-mythos-5-1m-mtp-q4-k-m", "efficiency": { "cold_start_seconds": 97.61175223300233, "cold_ttft_seconds": 0.0014789390043006279, "median_task_seconds": 252.813, "tasks_per_minute": 0.2373, "total_seconds": 252.813, "workload_seconds": 155.2012724169981, "workload_ttft_seconds": 0.03919271699851379 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 1, "core": 1, "strict": 1, "total": 1 }, "failure_counts": { "pass": 1 }, "manifest_path": "runs/20260812t015120-latency-empero-ai-qwythos-9b-claude-mythos-5-1m-gguf-qwythos-9b-claude-mythos-5-r1/manifest.json", "model": "empero-ai-qwythos-9b-claude-mythos-5-1m-gguf-qwythos-9b-claude-mythos-5-1m-mtp-q4-k-m", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 1, "evaluated_pass_rate": 100.0, "scheduled": 1, "scheduled_pass_rate": 100.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260812t015120-latency-empero-ai-qwythos-9b-claude-mythos-5-1m-gguf-qwythos-9b-claude-mythos-5-r1", "score": 100.0, "started_at": "2026-08-12T01:51:20.432340-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260812t015120-latency-empero-ai-qwythos-9b-claude-mythos-5-1m-gguf-qwythos-9b-claude-mythos-5-r1/summary.json", "task_results_path": "runs/20260812t015120-latency-empero-ai-qwythos-9b-claude-mythos-5-1m-gguf-qwythos-9b-claude-mythos-5-r1/task-results.jsonl" }, { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": {}, "display_name": "gemma-4-12b-obliterated", "efficiency": { "cold_start_seconds": 11.65942994600482, "cold_ttft_seconds": 0.004682411999965552, "median_task_seconds": 14.151, "tasks_per_minute": 4.24, "total_seconds": 14.151, "workload_seconds": 2.4915786320052575, "workload_ttft_seconds": 0.9312468170028296 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 1, "core": 1, "strict": 1, "total": 1 }, "failure_counts": { "pass": 1 }, "manifest_path": "runs/20260812t015120-latency-gemma-4-12b-obliterated-r1/manifest.json", "model": "gemma-4-12b-obliterated", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 1, "evaluated_pass_rate": 100.0, "scheduled": 1, "scheduled_pass_rate": 100.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260812t015120-latency-gemma-4-12b-obliterated-r1", "score": 100.0, "started_at": "2026-08-12T01:51:20.432340-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260812t015120-latency-gemma-4-12b-obliterated-r1/summary.json", "task_results_path": "runs/20260812t015120-latency-gemma-4-12b-obliterated-r1/task-results.jsonl" }, { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": {}, "display_name": "gemma-4-e4b-it", "efficiency": { "cold_start_seconds": 9.733592305005004, "cold_ttft_seconds": 0.005922943004406989, "median_task_seconds": 11.2863, "tasks_per_minute": 5.3162, "total_seconds": 11.2863, "workload_seconds": 1.552729579998413, "workload_ttft_seconds": 0.7556925419994514 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 1, "core": 1, "strict": 1, "total": 1 }, "failure_counts": { "pass": 1 }, "manifest_path": "runs/20260812t015120-latency-gemma-4-e4b-it-r1/manifest.json", "model": "gemma-4-e4b-it", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 1, "evaluated_pass_rate": 100.0, "scheduled": 1, "scheduled_pass_rate": 100.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260812t015120-latency-gemma-4-e4b-it-r1", "score": 100.0, "started_at": "2026-08-12T01:51:20.432340-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260812t015120-latency-gemma-4-e4b-it-r1/summary.json", "task_results_path": "runs/20260812t015120-latency-gemma-4-e4b-it-r1/task-results.jsonl" }, { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": {}, "display_name": "qwen3.6-27b-heretic-neo-code", "efficiency": { "cold_start_seconds": 35.27423781199468, "cold_ttft_seconds": 0.0023845429968787357, "median_task_seconds": 58.4755, "tasks_per_minute": 1.0261, "total_seconds": 58.4755, "workload_seconds": 23.201270860998193, "workload_ttft_seconds": 4.8704243960019085 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 1, "core": 1, "strict": 1, "total": 1 }, "failure_counts": { "pass": 1 }, "manifest_path": "runs/20260812t015120-latency-qwen3-6-27b-heretic-neo-code-r1/manifest.json", "model": "qwen3.6-27b-heretic-neo-code", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 1, "evaluated_pass_rate": 100.0, "scheduled": 1, "scheduled_pass_rate": 100.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260812t015120-latency-qwen3-6-27b-heretic-neo-code-r1", "score": 100.0, "started_at": "2026-08-12T01:51:20.432340-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260812t015120-latency-qwen3-6-27b-heretic-neo-code-r1/summary.json", "task_results_path": "runs/20260812t015120-latency-qwen3-6-27b-heretic-neo-code-r1/task-results.jsonl" }, { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": {}, "display_name": "qwen3.6-35b-hauhaucs-aggressive", "efficiency": { "cold_start_seconds": 29.923409277005703, "cold_ttft_seconds": 0.003349618003994692, "median_task_seconds": 32.048, "tasks_per_minute": 1.8722, "total_seconds": 32.048, "workload_seconds": 2.1245510640001157, "workload_ttft_seconds": 1.4488617530005286 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 1, "core": 1, "strict": 1, "total": 1 }, "failure_counts": { "pass": 1 }, "manifest_path": "runs/20260812t015120-latency-qwen3-6-35b-hauhaucs-aggressive-r1/manifest.json", "model": "qwen3.6-35b-hauhaucs-aggressive", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 1, "evaluated_pass_rate": 100.0, "scheduled": 1, "scheduled_pass_rate": 100.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260812t015120-latency-qwen3-6-35b-hauhaucs-aggressive-r1", "score": 100.0, "started_at": "2026-08-12T01:51:20.432340-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260812t015120-latency-qwen3-6-35b-hauhaucs-aggressive-r1/summary.json", "task_results_path": "runs/20260812t015120-latency-qwen3-6-35b-hauhaucs-aggressive-r1/task-results.jsonl" }, { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": {}, "display_name": "unsloth-qwen3-6-27b-mtp-gguf-qwen3-6-27b-ud-q4-k-xl", "efficiency": { "cold_start_seconds": 42.776272613999026, "cold_ttft_seconds": 0.00373033599316841, "median_task_seconds": 70.9961, "tasks_per_minute": 0.8451, "total_seconds": 70.9961, "workload_seconds": 28.219845717998396, "workload_ttft_seconds": 4.807563656999264 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 1, "core": 1, "strict": 1, "total": 1 }, "failure_counts": { "pass": 1 }, "manifest_path": "runs/20260812t015120-latency-unsloth-qwen3-6-27b-mtp-gguf-qwen3-6-27b-ud-q4-k-xl-r1/manifest.json", "model": "unsloth-qwen3-6-27b-mtp-gguf-qwen3-6-27b-ud-q4-k-xl", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 1, "evaluated_pass_rate": 100.0, "scheduled": 1, "scheduled_pass_rate": 100.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260812t015120-latency-unsloth-qwen3-6-27b-mtp-gguf-qwen3-6-27b-ud-q4-k-xl-r1", "score": 100.0, "started_at": "2026-08-12T01:51:20.432340-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260812t015120-latency-unsloth-qwen3-6-27b-mtp-gguf-qwen3-6-27b-ud-q4-k-xl-r1/summary.json", "task_results_path": "runs/20260812t015120-latency-unsloth-qwen3-6-27b-mtp-gguf-qwen3-6-27b-ud-q4-k-xl-r1/task-results.jsonl" }, { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": {}, "display_name": "unsloth-qwen3-6-35b-a3b-mtp-gguf-qwen3-6-35b-a3b-ud-q4-k-s", "efficiency": { "cold_start_seconds": 29.948987545001728, "cold_ttft_seconds": 0.0035129169991705567, "median_task_seconds": 32.3097, "tasks_per_minute": 1.857, "total_seconds": 32.3097, "workload_seconds": 2.360662939005124, "workload_ttft_seconds": 1.556977263004228 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 1, "core": 1, "strict": 1, "total": 1 }, "failure_counts": { "pass": 1 }, "manifest_path": "runs/20260812t015120-latency-unsloth-qwen3-6-35b-a3b-mtp-gguf-qwen3-6-35b-a3b-ud-q4-k-s-r1/manifest.json", "model": "unsloth-qwen3-6-35b-a3b-mtp-gguf-qwen3-6-35b-a3b-ud-q4-k-s", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 1, "evaluated_pass_rate": 100.0, "scheduled": 1, "scheduled_pass_rate": 100.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260812t015120-latency-unsloth-qwen3-6-35b-a3b-mtp-gguf-qwen3-6-35b-a3b-ud-q4-k-s-r1", "score": 100.0, "started_at": "2026-08-12T01:51:20.432340-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260812t015120-latency-unsloth-qwen3-6-35b-a3b-mtp-gguf-qwen3-6-35b-a3b-ud-q4-k-s-r1/summary.json", "task_results_path": "runs/20260812t015120-latency-unsloth-qwen3-6-35b-a3b-mtp-gguf-qwen3-6-35b-a3b-ud-q4-k-s-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": { "efficiency": 100.0, "english": 83.607, "grounding": 27.717, "outcome": 16.913, "safety": 92.918, "state": 39.614, "tool_use": 10.922 }, "display_name": "Cydonia 24B v4.3 Q4_K_M", "efficiency": { "median_task_seconds": 24.6922, "tasks_per_minute": 1.4671, "total_seconds": 858.8454 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 3, "core": 8, "strict": 3, "total": 21 }, "failure_counts": { "model_failure": 18, "pass": 3 }, "manifest_path": "runs/20260812t020233-assistant-cydonia-24b-v4-3-r1/manifest.json", "model": "cydonia-24b-v4.3", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 21, "evaluated_pass_rate": 14.286, "scheduled": 21, "scheduled_pass_rate": 14.286 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260812t020233-assistant-cydonia-24b-v4-3-r1", "score": 47.941, "started_at": "2026-08-12T02:02:33.983968-04:00", "status": "ok", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260812t020233-assistant-cydonia-24b-v4-3-r1/summary.json", "task_results_path": "runs/20260812t020233-assistant-cydonia-24b-v4-3-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": { "efficiency": 100.0, "english": 87.705, "grounding": 19.565, "outcome": 13.742, "safety": 92.918, "state": 39.614, "tool_use": 10.922 }, "display_name": "Dolphin Mistral 24B Venice Edition Q4_K_M", "efficiency": { "median_task_seconds": 1.9142, "tasks_per_minute": 19.3627, "total_seconds": 65.0736 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 2, "core": 8, "strict": 2, "total": 21 }, "failure_counts": { "model_failure": 19, "pass": 2 }, "manifest_path": "runs/20260812t020233-assistant-dolphin-mistral-24b-venice-r1/manifest.json", "model": "dolphin-mistral-24b-venice", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 21, "evaluated_pass_rate": 9.524, "scheduled": 21, "scheduled_pass_rate": 9.524 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260812t020233-assistant-dolphin-mistral-24b-venice-r1", "score": 46.989, "started_at": "2026-08-12T02:02:33.983968-04:00", "status": "ok", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260812t020233-assistant-dolphin-mistral-24b-venice-r1/summary.json", "task_results_path": "runs/20260812t020233-assistant-dolphin-mistral-24b-venice-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": { "efficiency": 78.571, "english": 83.607, "grounding": 49.457, "outcome": 37.844, "safety": 90.652, "state": 58.937, "tool_use": 85.437 }, "display_name": "Qwythos-9B-Claude-Mythos-5-1M-MTP-Q4_K_M", "efficiency": { "median_task_seconds": 107.6179, "tasks_per_minute": 0.4976, "total_seconds": 2291.1681 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 11, "core": 15, "strict": 11, "total": 21 }, "failure_counts": { "harness_error": 2, "model_failure": 8, "pass": 11 }, "manifest_path": "runs/20260812t020233-assistant-empero-ai-qwythos-9b-claude-mythos-5-1m-gguf-qwythos-9b-claude-mythos-5-r1/manifest.json", "model": "empero-ai-qwythos-9b-claude-mythos-5-1m-gguf-qwythos-9b-claude-mythos-5-1m-mtp-q4-k-m", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 19, "evaluated_pass_rate": 57.895, "scheduled": 21, "scheduled_pass_rate": 52.381 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260812t020233-assistant-empero-ai-qwythos-9b-claude-mythos-5-1m-gguf-qwythos-9b-claude-mythos-5-r1", "score": 68.792, "started_at": "2026-08-12T02:02:33.983968-04:00", "status": "partial", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260812t020233-assistant-empero-ai-qwythos-9b-claude-mythos-5-1m-gguf-qwythos-9b-claude-mythos-5-r1/summary.json", "task_results_path": "runs/20260812t020233-assistant-empero-ai-qwythos-9b-claude-mythos-5-1m-gguf-qwythos-9b-claude-mythos-5-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": { "efficiency": 88.095, "english": 59.016, "grounding": 35.87, "outcome": 18.393, "safety": 97.167, "state": 39.614, "tool_use": 68.204 }, "display_name": "Gemma 4 12B Obliterated Q4_K_M", "efficiency": { "median_task_seconds": 11.0058, "tasks_per_minute": 5.6594, "total_seconds": 222.6403 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 6, "core": 11, "strict": 6, "total": 21 }, "failure_counts": { "model_failure": 13, "pass": 6, "tool_protocol_failure": 2 }, "manifest_path": "runs/20260812t020233-assistant-gemma-4-12b-obliterated-r1/manifest.json", "model": "gemma-4-12b-obliterated", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 21, "evaluated_pass_rate": 28.571, "scheduled": 21, "scheduled_pass_rate": 28.571 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260812t020233-assistant-gemma-4-12b-obliterated-r1", "score": 57.036, "started_at": "2026-08-12T02:02:33.983968-04:00", "status": "ok", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260812t020233-assistant-gemma-4-12b-obliterated-r1/summary.json", "task_results_path": "runs/20260812t020233-assistant-gemma-4-12b-obliterated-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": { "efficiency": 90.476, "english": 91.803, "grounding": 46.739, "outcome": 34.038, "safety": 97.167, "state": 68.599, "tool_use": 64.563 }, "display_name": "Gemma 4 E4B IT UD-Q4_K_XL", "efficiency": { "median_task_seconds": 4.5186, "tasks_per_minute": 10.8379, "total_seconds": 110.7228 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 9, "core": 13, "strict": 9, "total": 21 }, "failure_counts": { "harness_error": 1, "model_failure": 10, "pass": 9, "tool_protocol_failure": 1 }, "manifest_path": "runs/20260812t020233-assistant-gemma-4-e4b-it-r1/manifest.json", "model": "gemma-4-e4b-it", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 20, "evaluated_pass_rate": 45.0, "scheduled": 21, "scheduled_pass_rate": 42.857 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260812t020233-assistant-gemma-4-e4b-it-r1", "score": 67.561, "started_at": "2026-08-12T02:02:33.983968-04:00", "status": "partial", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260812t020233-assistant-gemma-4-e4b-it-r1/summary.json", "task_results_path": "runs/20260812t020233-assistant-gemma-4-e4b-it-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": { "efficiency": 78.571, "english": 83.607, "grounding": 57.609, "outcome": 43.975, "safety": 94.901, "state": 80.676, "tool_use": 81.796 }, "display_name": "Qwen3.6 27B Heretic NEO CODE Q4_K_M", "efficiency": { "median_task_seconds": 173.5874, "tasks_per_minute": 0.2597, "total_seconds": 4158.3172 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 13, "core": 16, "strict": 12, "total": 21 }, "failure_counts": { "harness_error": 3, "model_failure": 6, "pass": 12 }, "manifest_path": "runs/20260812t020233-assistant-qwen3-6-27b-heretic-neo-code-r1/manifest.json", "model": "qwen3.6-27b-heretic-neo-code", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 18, "evaluated_pass_rate": 66.667, "scheduled": 21, "scheduled_pass_rate": 57.143 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260812t020233-assistant-qwen3-6-27b-heretic-neo-code-r1", "score": 73.083, "started_at": "2026-08-12T02:02:33.983968-04:00", "status": "partial", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260812t020233-assistant-qwen3-6-27b-heretic-neo-code-r1/summary.json", "task_results_path": "runs/20260812t020233-assistant-qwen3-6-27b-heretic-neo-code-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": { "efficiency": 64.286, "english": 79.508, "grounding": 38.587, "outcome": 43.975, "safety": 84.986, "state": 49.275, "tool_use": 85.437 }, "display_name": "Qwen3.6 35B A3B Uncensored HauhauCS Aggressive Q4_K_M", "efficiency": { "median_task_seconds": 9.0143, "tasks_per_minute": 4.7202, "total_seconds": 216.0944 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 9, "core": 14, "strict": 9, "total": 21 }, "failure_counts": { "harness_error": 4, "model_failure": 8, "pass": 9 }, "manifest_path": "runs/20260812t020233-assistant-qwen3-6-35b-hauhaucs-aggressive-r1/manifest.json", "model": "qwen3.6-35b-hauhaucs-aggressive", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 17, "evaluated_pass_rate": 52.941, "scheduled": 21, "scheduled_pass_rate": 42.857 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260812t020233-assistant-qwen3-6-35b-hauhaucs-aggressive-r1", "score": 65.436, "started_at": "2026-08-12T02:02:33.983968-04:00", "status": "partial", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260812t020233-assistant-qwen3-6-35b-hauhaucs-aggressive-r1/summary.json", "task_results_path": "runs/20260812t020233-assistant-qwen3-6-35b-hauhaucs-aggressive-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": { "efficiency": 78.571, "english": 83.607, "grounding": 65.761, "outcome": 61.311, "safety": 100.0, "state": 63.768, "tool_use": 83.981 }, "display_name": "Qwen3.6-27B-UD-Q4_K_XL", "efficiency": { "median_task_seconds": 258.8873, "tasks_per_minute": 0.2193, "total_seconds": 5470.9727 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 13, "core": 16, "strict": 13, "total": 21 }, "failure_counts": { "harness_error": 1, "model_failure": 7, "pass": 13 }, "manifest_path": "runs/20260812t020233-assistant-unsloth-qwen3-6-27b-mtp-gguf-qwen3-6-27b-ud-q4-k-xl-r1/manifest.json", "model": "unsloth-qwen3-6-27b-mtp-gguf-qwen3-6-27b-ud-q4-k-xl", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 20, "evaluated_pass_rate": 65.0, "scheduled": 21, "scheduled_pass_rate": 61.905 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260812t020233-assistant-unsloth-qwen3-6-27b-mtp-gguf-qwen3-6-27b-ud-q4-k-xl-r1", "score": 77.268, "started_at": "2026-08-12T02:02:33.983968-04:00", "status": "partial", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260812t020233-assistant-unsloth-qwen3-6-27b-mtp-gguf-qwen3-6-27b-ud-q4-k-xl-r1/summary.json", "task_results_path": "runs/20260812t020233-assistant-unsloth-qwen3-6-27b-mtp-gguf-qwen3-6-27b-ud-q4-k-xl-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result.", "Imported from a historical cohort result; the bundle is normalized evidence, not a rerun." ], "cohort": "titan-local", "dimensions": { "efficiency": 76.19, "english": 81.557, "grounding": 60.326, "outcome": 41.86, "safety": 94.901, "state": 63.768, "tool_use": 85.437 }, "display_name": "Qwen3.6-35B-A3B-UD-Q4_K_S", "efficiency": { "median_task_seconds": 11.4967, "tasks_per_minute": 5.0627, "total_seconds": 237.0257 }, "endpoint_class": "llama.cpp", "exactness": { "contract": 11, "core": 15, "strict": 11, "total": 21 }, "failure_counts": { "harness_error": 1, "model_failure": 9, "pass": 11 }, "manifest_path": "runs/20260812t020233-assistant-unsloth-qwen3-6-35b-a3b-mtp-gguf-qwen3-6-35b-a3b-ud-q4-k-s-r1/manifest.json", "model": "unsloth-qwen3-6-35b-a3b-mtp-gguf-qwen3-6-35b-a3b-ud-q4-k-s", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": null, "reliability": { "evaluated": 20, "evaluated_pass_rate": 55.0, "scheduled": 21, "scheduled_pass_rate": 52.381 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "7362428bf3ca09c95423d2c2c1c5dd11bea534d8", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260812t020233-assistant-unsloth-qwen3-6-35b-a3b-mtp-gguf-qwen3-6-35b-a3b-ud-q4-k-s-r1", "score": 71.354, "started_at": "2026-08-12T02:02:33.983968-04:00", "status": "partial", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260812t020233-assistant-unsloth-qwen3-6-35b-a3b-mtp-gguf-qwen3-6-35b-a3b-ud-q4-k-s-r1/summary.json", "task_results_path": "runs/20260812t020233-assistant-unsloth-qwen3-6-35b-a3b-mtp-gguf-qwen3-6-35b-a3b-ud-q4-k-s-r1/task-results.jsonl" }, { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent." ], "cohort": "titan-local", "dimensions": {}, "display_name": "hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-uncensored-hauhaucs-aggressive-q4-k-p", "efficiency": { "cold_start_seconds": 52.09204200201202, "cold_ttft_seconds": 0.0017128739855252206, "median_task_seconds": 54.7107, "tasks_per_minute": 1.0967, "total_seconds": 54.7107, "workload_seconds": 2.6186941580381244, "workload_ttft_seconds": 2.1793475979939103 }, "endpoint_class": "llama.cpp/OpenAI-compatible", "exactness": { "contract": 1, "core": 1, "strict": 1, "total": 1 }, "failure_counts": { "pass": 1 }, "manifest_path": "runs/20260906t210451-latency-hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-r1/manifest.json", "model": "hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-uncensored-hauhaucs-aggressive-q4-k-p", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": 0.0, "reliability": { "evaluated": 1, "evaluated_pass_rate": 100.0, "scheduled": 1, "scheduled_pass_rate": 100.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "6b41b1d2ee7c0f681837e03ec4f9b2e2f6b1339a", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260906t210451-latency-hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-r1", "score": 100.0, "started_at": "2026-09-06T21:04:51.741961-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260906t210451-latency-hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-r1/summary.json", "task_results_path": "runs/20260906t210451-latency-hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result." ], "cohort": "titan-local", "dimensions": { "efficiency": 71.429, "english": 69.262, "grounding": 63.043, "outcome": 42.918, "safety": 100.0, "state": 63.768, "tool_use": 86.893 }, "display_name": "Qwen3.8 27B Uncensored HauhauCS Aggressive Q4_K_P — RTX 3090 MTP", "efficiency": { "median_task_seconds": 7.1695, "tasks_per_minute": 5.4802, "total_seconds": 164.2264 }, "endpoint_class": "llama.cpp/OpenAI-compatible", "exactness": { "contract": 10, "core": 14, "strict": 10, "total": 21 }, "failure_counts": { "harness_error": 6, "model_failure": 5, "pass": 10 }, "manifest_path": "runs/20260906t210610-assistant-hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-r1/manifest.json", "model": "hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-uncensored-hauhaucs-aggressive-q4-k-p", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": 0.0, "reliability": { "evaluated": 15, "evaluated_pass_rate": 66.667, "scheduled": 21, "scheduled_pass_rate": 47.619 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "6b41b1d2ee7c0f681837e03ec4f9b2e2f6b1339a", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260906t210610-assistant-hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-r1", "score": 70.97, "started_at": "2026-09-06T21:06:10.019751-04:00", "status": "partial", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260906t210610-assistant-hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-r1/summary.json", "task_results_path": "runs/20260906t210610-assistant-hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-r1/task-results.jsonl" }, { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent." ], "cohort": "titan-local", "dimensions": {}, "display_name": "jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m", "efficiency": { "cold_start_seconds": 35.12104332103627, "cold_ttft_seconds": 0.0016259429976344109, "median_task_seconds": 37.4871, "tasks_per_minute": 1.6006, "total_seconds": 37.4871, "workload_seconds": 2.3660284419893287, "workload_ttft_seconds": 1.9589812509948388 }, "endpoint_class": "llama.cpp/OpenAI-compatible", "exactness": { "contract": 1, "core": 1, "strict": 1, "total": 1 }, "failure_counts": { "pass": 1 }, "manifest_path": "runs/20260906t211419-latency-jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-r1/manifest.json", "model": "jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": 0.0, "reliability": { "evaluated": 1, "evaluated_pass_rate": 100.0, "scheduled": 1, "scheduled_pass_rate": 100.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "04f0f5d7825fe81f700b1f3c52966022a07a4c85", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260906t211419-latency-jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-r1", "score": 100.0, "started_at": "2026-09-06T21:14:19.975824-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260906t211419-latency-jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-r1/summary.json", "task_results_path": "runs/20260906t211419-latency-jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result." ], "cohort": "titan-local", "dimensions": { "efficiency": 73.81, "english": 73.361, "grounding": 63.043, "outcome": 49.26, "safety": 92.068, "state": 80.676, "tool_use": 91.748 }, "display_name": "Qwen3.8-27B Uncensored Q4_K_M — RTX 3090 MTP", "efficiency": { "median_task_seconds": 6.8359, "tasks_per_minute": 5.7006, "total_seconds": 157.8782 }, "endpoint_class": "llama.cpp/OpenAI-compatible", "exactness": { "contract": 11, "core": 16, "strict": 11, "total": 21 }, "failure_counts": { "harness_error": 6, "model_failure": 4, "pass": 11 }, "manifest_path": "runs/20260906t211514-assistant-jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-r1/manifest.json", "model": "jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": 0.0, "reliability": { "evaluated": 15, "evaluated_pass_rate": 73.333, "scheduled": 21, "scheduled_pass_rate": 52.381 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "04f0f5d7825fe81f700b1f3c52966022a07a4c85", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260906t211514-assistant-jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-r1", "score": 74.549, "started_at": "2026-09-06T21:15:14.814539-04:00", "status": "partial", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260906t211514-assistant-jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-r1/summary.json", "task_results_path": "runs/20260906t211514-assistant-jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-r1/task-results.jsonl" }, { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent." ], "cohort": "titan-local", "dimensions": {}, "display_name": "orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m", "efficiency": { "cold_start_seconds": 31.96425466798246, "cold_ttft_seconds": 0.007055675960145891, "median_task_seconds": 34.2603, "tasks_per_minute": 1.7513, "total_seconds": 34.2603, "workload_seconds": 2.296047699986957, "workload_ttft_seconds": 1.8848411400103942 }, "endpoint_class": "llama.cpp/OpenAI-compatible", "exactness": { "contract": 1, "core": 1, "strict": 1, "total": 1 }, "failure_counts": { "pass": 1 }, "manifest_path": "runs/20260906t212131-latency-orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m-r1/manifest.json", "model": "orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": 0.0, "reliability": { "evaluated": 1, "evaluated_pass_rate": 100.0, "scheduled": 1, "scheduled_pass_rate": 100.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "fe22f57391fc8ab837b86140919dae84c131abd6", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260906t212131-latency-orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m-r1", "score": 100.0, "started_at": "2026-09-06T21:21:31.365986-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260906t212131-latency-orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m-r1/summary.json", "task_results_path": "runs/20260906t212131-latency-orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result." ], "cohort": "titan-local", "dimensions": { "efficiency": 64.286, "english": 73.361, "grounding": 71.196, "outcome": 51.374, "safety": 97.167, "state": 63.768, "tool_use": 86.893 }, "display_name": "Qwen3.8 27B Uncensored OrcaRouter Q4_K_M — RTX 3090 MTP", "efficiency": { "median_task_seconds": 9.1904, "tasks_per_minute": 5.3611, "total_seconds": 212.6447 }, "endpoint_class": "llama.cpp/OpenAI-compatible", "exactness": { "contract": 12, "core": 16, "strict": 12, "total": 21 }, "failure_counts": { "harness_error": 2, "model_failure": 7, "pass": 12 }, "manifest_path": "runs/20260906t212223-assistant-orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m-r1/manifest.json", "model": "orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": 0.0, "reliability": { "evaluated": 19, "evaluated_pass_rate": 63.158, "scheduled": 21, "scheduled_pass_rate": 57.143 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "fe22f57391fc8ab837b86140919dae84c131abd6", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260906t212223-assistant-orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m-r1", "score": 72.963, "started_at": "2026-09-06T21:22:23.313691-04:00", "status": "partial", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260906t212223-assistant-orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m-r1/summary.json", "task_results_path": "runs/20260906t212223-assistant-orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m-r1/task-results.jsonl" }, { "caveats": [ "Cold-load control is available only for local lifecycle-managed runtimes; hosted-provider latency is directional, not equivalent." ], "cohort": "titan-local", "dimensions": {}, "display_name": "unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl", "efficiency": { "cold_start_seconds": 35.0803113830043, "cold_ttft_seconds": 0.001716608996503055, "median_task_seconds": 37.4224, "tasks_per_minute": 1.6033, "total_seconds": 37.4224, "workload_seconds": 2.3421120359562337, "workload_ttft_seconds": 1.9254901659442112 }, "endpoint_class": "llama.cpp/OpenAI-compatible", "exactness": { "contract": 1, "core": 1, "strict": 1, "total": 1 }, "failure_counts": { "pass": 1 }, "manifest_path": "runs/20260906t213243-latency-unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl-r1/manifest.json", "model": "unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": 0.0, "reliability": { "evaluated": 1, "evaluated_pass_rate": 100.0, "scheduled": 1, "scheduled_pass_rate": 100.0 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": null, "prompt_sha256": "1424e431a5061a12479bc53fcec60050baa8009840d27007c9109813c29fcbbb", "runner_commit": "6b1792b1eac3afefdd409b123770c533f09dffc7", "runner_dirty": true, "seed_document_sha256": "56994618c221c496c6c7e812d4a8605158836880e554c9ff23ae0b09e52060c4", "tasks_sha256": "f417029b6283e23974e4a1f58a8b0f44294af6ba28e421b272dc2edf752f0aea", "tool_schema_sha256": "ff9794396f5cd5d59d82b27fad8325f2fac2df7a647eda762261a1c02cad3e55" }, "run_id": "20260906t213243-latency-unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl-r1", "score": 100.0, "started_at": "2026-09-06T21:32:43.538300-04:00", "status": "ok", "suite": "latency", "suite_version": "1.0", "summary_path": "runs/20260906t213243-latency-unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl-r1/summary.json", "task_results_path": "runs/20260906t213243-latency-unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl-r1/task-results.jsonl" }, { "caveats": [ "Core and contract exactness are derived from legacy assistant-v1 score thresholds (50 and 70); strict preserves the original task pass result." ], "cohort": "titan-local", "dimensions": { "efficiency": 66.667, "english": 77.459, "grounding": 71.196, "outcome": 52.854, "safety": 97.167, "state": 63.768, "tool_use": 86.893 }, "display_name": "Qwen3.8 27B UD Q4_K_XL (tuned)", "efficiency": { "median_task_seconds": 8.3916, "tasks_per_minute": 6.1613, "total_seconds": 175.2869 }, "endpoint_class": "llama.cpp/OpenAI-compatible", "exactness": { "contract": 12, "core": 16, "strict": 12, "total": 21 }, "failure_counts": { "harness_error": 3, "model_failure": 6, "pass": 12 }, "manifest_path": "runs/20260906t213340-assistant-unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl-r1/manifest.json", "model": "unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": 0.0, "reliability": { "evaluated": 18, "evaluated_pass_rate": 66.667, "scheduled": 21, "scheduled_pass_rate": 57.143 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ea2601bcb637a9c66563e91015f15a007a0a06f7d88532ec218c8c178903efb9", "prompt_sha256": "be032d2eb70ed8bd9486cdb34502b0aecc7d108f2912c2acc53567bacbcfbc27", "runner_commit": "6b1792b1eac3afefdd409b123770c533f09dffc7", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "907017aa0d29639967cd0c0702764b73e7cc8ebbf254625fcc07b63e13b4428e", "tool_schema_sha256": "5f50178d6b57bf3e4b395d941772f3cba613a17d065a7c9161a09c5617ca0d17" }, "run_id": "20260906t213340-assistant-unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl-r1", "score": 74.101, "started_at": "2026-09-06T21:33:40.027386-04:00", "status": "partial", "suite": "assistant", "suite_version": "1.0", "summary_path": "runs/20260906t213340-assistant-unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl-r1/summary.json", "task_results_path": "runs/20260906t213340-assistant-unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl-r1/task-results.jsonl" }, { "caveats": [], "cohort": "titan-local", "dimensions": { "efficiency": 100.0, "english": 100.0, "factuality": 66.667, "outcome": 87.5, "safety": 66.667, "sourcing": 100.0, "structure": 50.0 }, "display_name": "Qwen3.8 27B Uncensored HauhauCS Aggressive Q4_K_P — RTX 3090 MTP", "efficiency": { "median_task_seconds": 7.1874, "tasks_per_minute": 8.3984, "total_seconds": 64.2976 }, "endpoint_class": "llama.cpp/OpenAI-compatible", "exactness": { "contract": 4, "core": 5, "strict": 3, "total": 9 }, "failure_counts": { "model_failure": 6, "pass": 3 }, "manifest_path": "runs/20260907t011004-editorial-hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-r1/manifest.json", "model": "hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-uncensored-hauhaucs-aggressive-q4-k-p", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": 0.0, "reliability": { "evaluated": 9, "evaluated_pass_rate": 33.333, "scheduled": 9, "scheduled_pass_rate": 33.333 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ce0a515b76b4550ed87f12183ef2f89eeedfb0bdd13dc4151253fdfea35f53ed", "prompt_sha256": "8db0b61683a69b744028f7fc249c67476d815233ed537a868c57e0aef6586ccf", "runner_commit": "6b41b1d2ee7c0f681837e03ec4f9b2e2f6b1339a", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "d6aa0c592f3ca85be4fb193e41aa88bbde928532f2b71b4cdd965c2bec1e1d68", "tool_schema_sha256": null }, "run_id": "20260907t011004-editorial-hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-r1", "score": 76.111, "started_at": "2026-09-07T01:10:04.163438+00:00", "status": "ok", "suite": "editorial", "suite_version": "editorial-v1", "summary_path": "runs/20260907t011004-editorial-hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-r1/summary.json", "task_results_path": "runs/20260907t011004-editorial-hauhaucs-qwen3-8-27b-uncensored-hauhaucs-aggressive-mtp-gguf-qwen3-8-27b-r1/task-results.jsonl" }, { "caveats": [], "cohort": "titan-local", "dimensions": { "efficiency": 66.667, "english": 100.0, "factuality": 100.0, "outcome": 100.0, "safety": 66.667, "sourcing": 100.0, "structure": 50.0 }, "display_name": "Qwen3.8-27B Uncensored Q4_K_M — RTX 3090 MTP", "efficiency": { "median_task_seconds": 7.7556, "tasks_per_minute": 8.4275, "total_seconds": 64.0758 }, "endpoint_class": "llama.cpp/OpenAI-compatible", "exactness": { "contract": 7, "core": 9, "strict": 4, "total": 9 }, "failure_counts": { "model_failure": 5, "pass": 4 }, "manifest_path": "runs/20260907t011837-editorial-jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-r1/manifest.json", "model": "jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": 0.0, "reliability": { "evaluated": 9, "evaluated_pass_rate": 44.444, "scheduled": 9, "scheduled_pass_rate": 44.444 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ce0a515b76b4550ed87f12183ef2f89eeedfb0bdd13dc4151253fdfea35f53ed", "prompt_sha256": "8db0b61683a69b744028f7fc249c67476d815233ed537a868c57e0aef6586ccf", "runner_commit": "04f0f5d7825fe81f700b1f3c52966022a07a4c85", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "d6aa0c592f3ca85be4fb193e41aa88bbde928532f2b71b4cdd965c2bec1e1d68", "tool_schema_sha256": null }, "run_id": "20260907t011837-editorial-jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-r1", "score": 89.444, "started_at": "2026-09-07T01:18:37.682025+00:00", "status": "ok", "suite": "editorial", "suite_version": "editorial-v1", "summary_path": "runs/20260907t011837-editorial-jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-r1/summary.json", "task_results_path": "runs/20260907t011837-editorial-jonathancoletti-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-r1/task-results.jsonl" }, { "caveats": [], "cohort": "titan-local", "dimensions": { "efficiency": 100.0, "english": 100.0, "factuality": 88.889, "outcome": 75.0, "safety": 66.667, "sourcing": 100.0, "structure": 50.0 }, "display_name": "Qwen3.8 27B Uncensored OrcaRouter Q4_K_M — RTX 3090 MTP", "efficiency": { "median_task_seconds": 8.7499, "tasks_per_minute": 8.7156, "total_seconds": 61.9578 }, "endpoint_class": "llama.cpp/OpenAI-compatible", "exactness": { "contract": 5, "core": 6, "strict": 4, "total": 9 }, "failure_counts": { "model_failure": 5, "pass": 4 }, "manifest_path": "runs/20260907t012701-editorial-orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m-r1/manifest.json", "model": "orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": 0.0, "reliability": { "evaluated": 9, "evaluated_pass_rate": 44.444, "scheduled": 9, "scheduled_pass_rate": 44.444 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ce0a515b76b4550ed87f12183ef2f89eeedfb0bdd13dc4151253fdfea35f53ed", "prompt_sha256": "8db0b61683a69b744028f7fc249c67476d815233ed537a868c57e0aef6586ccf", "runner_commit": "fe22f57391fc8ab837b86140919dae84c131abd6", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "d6aa0c592f3ca85be4fb193e41aa88bbde928532f2b71b4cdd965c2bec1e1d68", "tool_schema_sha256": null }, "run_id": "20260907t012701-editorial-orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m-r1", "score": 81.667, "started_at": "2026-09-07T01:27:01.711567+00:00", "status": "ok", "suite": "editorial", "suite_version": "editorial-v1", "summary_path": "runs/20260907t012701-editorial-orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m-r1/summary.json", "task_results_path": "runs/20260907t012701-editorial-orcarouter-qwen3-8-27b-uncensored-gguf-qwen3-8-27b-uncensored-q4-k-m-r1/task-results.jsonl" }, { "caveats": [], "cohort": "titan-local", "dimensions": { "efficiency": 100.0, "english": 100.0, "factuality": 77.778, "outcome": 62.5, "safety": 66.667, "sourcing": 100.0, "structure": 50.0 }, "display_name": "Qwen3.8 27B UD Q4_K_XL (tuned)", "efficiency": { "median_task_seconds": 8.515, "tasks_per_minute": 8.8623, "total_seconds": 60.9323 }, "endpoint_class": "llama.cpp/OpenAI-compatible", "exactness": { "contract": 4, "core": 4, "strict": 3, "total": 9 }, "failure_counts": { "model_failure": 6, "pass": 3 }, "manifest_path": "runs/20260907t013750-editorial-unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl-r1/manifest.json", "model": "unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl", "provider": "Titan", "provider_kind": "local", "provider_reported_cost_usd": 0.0, "reliability": { "evaluated": 9, "evaluated_pass_rate": 33.333, "scheduled": 9, "scheduled_pass_rate": 33.333 }, "replicate": { "count": 1, "index": 1, "seed": 42 }, "reproducibility": { "fixture_sha256": "ce0a515b76b4550ed87f12183ef2f89eeedfb0bdd13dc4151253fdfea35f53ed", "prompt_sha256": "8db0b61683a69b744028f7fc249c67476d815233ed537a868c57e0aef6586ccf", "runner_commit": "6b1792b1eac3afefdd409b123770c533f09dffc7", "runner_dirty": true, "seed_document_sha256": "0dccd3175f37e10e978c3937f5aed91d2c482652f99a2842c4d18a5ab7713a44", "tasks_sha256": "d6aa0c592f3ca85be4fb193e41aa88bbde928532f2b71b4cdd965c2bec1e1d68", "tool_schema_sha256": null }, "run_id": "20260907t013750-editorial-unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl-r1", "score": 74.444, "started_at": "2026-09-07T01:37:50.911007+00:00", "status": "ok", "suite": "editorial", "suite_version": "editorial-v1", "summary_path": "runs/20260907t013750-editorial-unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl-r1/summary.json", "task_results_path": "runs/20260907t013750-editorial-unsloth-qwen3-8-27b-gguf-qwen3-8-27b-ud-q4-k-xl-r1/task-results.jsonl" } ], "schema_version": 1 }