MMLU-Pro
- What does it measure?
- Artificial Analysis now classifies this as 'legacy' and rarely measures new models, so recent ones may be missing. A harder version of MMLU, the exam built from university-level subjects. It tests knowledge and reasoning across a broad range of fields — law, medicine, engineering, history and more.
- Who evaluates it, and how?
- Researchers at the University of Waterloo and elsewhere expanded MMLU from four to ten answer options, removed questions that were too easy or flawed, and rebuilt it into about 12,000 questions across 14 fields. Scored as accuracy; Artificial Analysis runs it independently.
- Reading the score
- 0–100%. With ten options, guessing gives 10%. It is designed to score 16–33 points lower than the original MMLU, so do not compare it directly with MMLU numbers elsewhere.
Rank
Top score89.5%
Models tested28
Scoring basisHigher is better
Last updated2026.11.21
Model release date
How high is its intelligence index across knowledge work, coding and science?AA Intelligence Index v5
Claude Fable 5.2🥇
Claude Fable 5.2🥇AA Intelligence Index v5
Claude Fable 5.2🥇Which AI gives the answers people prefer most?Arena Text
Gemini 4 Carbon🥇
Gemini 4 Carbon🥇Arena Text
Gemini 4 Carbon🥇Can it solve problems it never trained on?LiveBench 2.0
Claude Fable 5.2🥇
Claude Fable 5.2🥇LiveBench 2.0
Claude Fable 5.2🥇
Does it have broad knowledge across expert fields?
MMLU-Pro
- 🥇AnthropicClaude Opus 4.589.5%
- 🥈GoogleGemini 3 Flash89.0%
- 🥉AnthropicClaude Opus 4.188.0%
- #4AnthropicClaude Sonnet 4.587.5%
- #5AnthropicClaude Opus 487.3%
- #6OpenAIGPT-587.1%
- #7DeepSeekDeepSeek V3.286.2%
- #7GoogleGemini 2.5 Pro86.2%
- #7UpstageSolar Pro 486.2%
- #10SpaceXSIGrok 4.1 Fast (Reasoning)85.4%
- #11AnthropicClaude Sonnet 484.2%
- #12LG AI ResearchK-EXAONE83.8%
- #13OpenAIGPT-5 Mini83.7%
- #14LG AI ResearchK-EXAONE 2.083.5%
- #15GoogleGemini 2.5 Flash83.2%
- #16BaiduERNIE 5.0 Thinking83.0%
- #17AmazonNova 2 Lite81.8%
- #18MetaLlama 4 Maverick80.9%
- #19GoogleGemini 2.5 Flash Lite80.8%
- #19OpenAIGPT OSS 120B80.8%
- #21OpenAIGPT-4.180.6%
- #22AnthropicClaude Haiku 4.580.0%
- #23NVIDIANemotron 3 Nano 30B A3B79.4%
- #24OpenAIGPT-5 Nano78.0%
- #25MetaLlama 4 Scout75.2%
- #26SpaceXSIGrok 4.1 Fast74.3%
- #27PerplexitySonar68.9%
- #28Mistral AIMistral Small 441.9%
Half of models ≤ 83.3%
59 of the 59 AIs released in the last 60 days have no score here yet · Gemini 4 Neon, GPT-6.2 Astra, Microsoft-Decision-1.1 and 56 more
This score comes from a fixed, predefined evaluation, so it may differ from what you get with your own question.
Scores labelled 'Self-reported' were published by the AI maker itself, not measured by a third party.