{"version": 1, "date": "2026-11-04", "taskSet": "2026-10-29.1", "tasks": 25, "runsPerTask": 3, "temperature": 0.0, "previousDate": "2026-10-29", "finding": "GPT-6 Luna, GPT-6.1 Terra and GPT-6.1 Sol each answered 25 of 25 checkable tasks correctly (100%), so these questions no longer tell them apart.", "categories": ["dates-units", "extraction", "instructions", "knowledge", "reasoning"], "models": [{"model": "gpt-6-luna", "label": "GPT-6 Luna", "where": "OpenAI API", "accuracy": 100.0, "accuracyChange": null, "correct": 25, "errors": 0, "latencyP50Seconds": 1.277, "costPerRunUsd": 0.000509, "costPer100TasksUsd": 0.002035, "tokens": 5078, "consistency": 84.0, "byCategory": {"dates-units": 100.0, "extraction": 100.0, "instructions": 100.0, "knowledge": 100.0, "reasoning": 100.0}, "priceSource": "openrouter.ai/api/v1/models read 2026-11-04 (list prices; recheck before publishing)"}, {"model": "gpt-6.1-terra", "label": "GPT-6.1 Terra", "where": "OpenAI API", "accuracy": 100.0, "accuracyChange": null, "correct": 25, "errors": 0, "latencyP50Seconds": 1.322, "costPerRunUsd": 0.003586, "costPer100TasksUsd": 0.014343, "tokens": 4177, "consistency": 92.0, "byCategory": {"dates-units": 100.0, "extraction": 100.0, "instructions": 100.0, "knowledge": 100.0, "reasoning": 100.0}, "priceSource": "openrouter.ai/api/v1/models read 2026-11-04 (list prices; recheck before publishing)"}, {"model": "gpt-6.1-sol", "label": "GPT-6.1 Sol", "where": "OpenAI API", "accuracy": 100.0, "accuracyChange": null, "correct": 25, "errors": 0, "latencyP50Seconds": 1.797, "costPerRunUsd": 0.006121, "costPer100TasksUsd": 0.024485, "tokens": 3862, "consistency": 96.0, "byCategory": {"dates-units": 100.0, "extraction": 100.0, "instructions": 100.0, "knowledge": 100.0, "reasoning": 100.0}, "priceSource": "openrouter.ai/api/v1/models read 2026-11-04 (list prices; recheck before publishing)"}, {"model": "gpt-4.1", "label": "GPT-4.1", "where": "OpenAI API", "accuracy": 96.0, "accuracyChange": 0.0, "correct": 24, "errors": 0, "latencyP50Seconds": 0.86, "costPerRunUsd": 0.002834, "costPer100TasksUsd": 0.011336, "tokens": 3018, "consistency": 92.0, "byCategory": {"dates-units": 75.0, "extraction": 100.0, "instructions": 100.0, "knowledge": 100.0, "reasoning": 100.0}, "priceSource": "openai.com/api/pricing (list prices; recheck before publishing)"}, {"model": "gpt-4o-mini", "label": "GPT-4o mini", "where": "OpenAI API", "accuracy": 92.0, "accuracyChange": 0.0, "correct": 23, "errors": 0, "latencyP50Seconds": 0.811, "costPerRunUsd": 0.000218, "costPer100TasksUsd": 0.000872, "tokens": 3045, "consistency": 100.0, "byCategory": {"dates-units": 75.0, "extraction": 100.0, "instructions": 80.0, "knowledge": 100.0, "reasoning": 100.0}, "priceSource": "openai.com/api/pricing (list prices; recheck before publishing)"}, {"model": "gpt-4.1-mini", "label": "GPT-4.1 mini", "where": "OpenAI API", "accuracy": 88.0, "accuracyChange": 0.0, "correct": 22, "errors": 0, "latencyP50Seconds": 0.786, "costPerRunUsd": 0.000575, "costPer100TasksUsd": 0.002299, "tokens": 3033, "consistency": 96.0, "byCategory": {"dates-units": 75.0, "extraction": 100.0, "instructions": 60.0, "knowledge": 100.0, "reasoning": 100.0}, "priceSource": "openai.com/api/pricing (list prices; recheck before publishing)"}], "notReached": [], "missedByAll": [], "sources": {"results": "C:\\Users\\AntiKali\\Documents\\Harness\\ModelBench\\2026-11-04\\results.csv", "methodology": "C:\\Users\\AntiKali\\Documents\\Harness\\ModelBench\\2026-11-04\\methodology.md"}, "harder": {"date": "2026-11-04", "taskSet": "2026-10-29.leap1", "tasks": 7, "runsPerTask": 3, "models": [{"model": "gpt-6.1-terra", "label": "GPT-6.1 Terra", "accuracy": 100.0, "byCategory": {"control": 100.0, "trap": 100.0}, "latencyP50Seconds": 1.576, "costPer100TasksUsd": 0.039971428571428574, "errors": 0}, {"model": "gpt-6.1-sol", "label": "GPT-6.1 Sol", "accuracy": 100.0, "byCategory": {"control": 100.0, "trap": 100.0}, "latencyP50Seconds": 2.234, "costPer100TasksUsd": 0.0648, "errors": 0}, {"model": "gpt-6-luna", "label": "GPT-6 Luna", "accuracy": 85.7, "byCategory": {"control": 83.3, "trap": 86.7}, "latencyP50Seconds": 2.124, "costPer100TasksUsd": 0.007613809523809524, "errors": 0}, {"model": "gpt-4.1", "label": "GPT-4.1", "accuracy": 66.7, "byCategory": {"control": 100.0, "trap": 53.3}, "latencyP50Seconds": 0.928, "costPer100TasksUsd": 0.05657142857142856, "errors": 0}], "categories": ["control", "trap"]}}