evals.report
BenchmarksLabsCompareRun guidesIn the wild
BenchmarksReasoning

AA-Omniscience: Knowledge and Hallucination Benchmark

A factuality and knowledge benchmark of 6,000 questions across 42 economically relevant topics in six domains, scoring models on the AA-Omniscience Index (-100 to 100) that rewards correct answers, penalizes hallucinations, and applies no penalty for abstaining.

ReasoningAA-Omniscience IndexHigher is better
ScoresAboutRun this benchmark

What is AA-Omniscience: Knowledge and Hallucination Benchmark?

A factuality and knowledge benchmark of 6,000 questions across 42 economically relevant topics in six domains, scoring models on the AA-Omniscience Index (-100 to 100) that rewards correct answers, penalizes hallucinations, and applies no penalty for abstaining. evals.report tracks reported AA-Omniscience: Knowledge and Hallucination Benchmark scores with the model, source, status, date, and run caveats attached — official leaderboard scores, vendor-reported launches, and clearly labeled community runs.

Top reported AA-Omniscience: Knowledge and Hallucination Benchmark score: Claude Fable 5.1 — 43 (AA-Omniscience Index).

ModelLabScore↓Source modelStatusDate
Claude Fable 5.1Anthropic43Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback)OfficialSep 1, 2026Details
Gemini 3.1 Pro PreviewGoogle DeepMind33—OfficialFeb 19, 2026Details
Claude Opus 4.8Anthropic27—OfficialMay 28, 2026Details
Claude Opus 4.7Anthropic26—OfficialApr 16, 2026Details
Gemini 3.5 FlashGoogle DeepMind23—OfficialMay 19, 2026Details
GPT-5.5OpenAI20—OfficialApr 23, 2026Details
Grok 4.3xAI18—OfficialApr 17, 2026Details
Claude Sonnet 4.6Anthropic12—OfficialFeb 17, 2026Details
Kimi K2.6OpenMoonshot AI6—OfficialApr 20, 2026Details
GPT-5.4OpenAI6—OfficialMar 5, 2026Details
Muse SparkMeta4—OfficialApr 8, 2026Details
MiMo-V2.5-ProOpenXiaomi4—OfficialApr 22, 2026Details
InklingOpenThinking Machines Lab2.1InklingVerifiedJul 15, 2026Details
GLM-5.1OpenZ.ai2—OfficialApr 7, 2026Details
MiniMax M2.7OpenMiniMax1—OfficialMar 18, 2026Details
Claude Haiku 4.5Anthropic-4—OfficialOct 15, 2025Details
DeepSeek V4 ProOpenDeepSeek-10—OfficialApr 24, 2026Details
Llama 3.1 405BOpenMeta-17—OfficialJul 23, 2024Details
DeepSeek V4 FlashOpenDeepSeek-23—OfficialApr 24, 2026Details
Qwen3.5-397B-A17BOpenAlibaba / Qwen-30—OfficialFeb 16, 2026Details
Mistral Medium 3.5Mistral AI-36—OfficialApr 28, 2026Details
NVIDIA Nemotron 3 Super 120B-A12BOpenNVIDIA-42—OfficialMar 10, 2026Details
GPT-OSS-120BOpenOpenAI-50—OfficialAug 5, 2025Details

Each row reports the model’s AA-Omniscience Index on AA-Omniscience: Knowledge and Hallucination Benchmark. Click a row for the full run context.