MMMU
MMMU (Massive Multi-discipline Multimodal Understanding) is a benchmark designed to evaluate multimodal models on college-level subject knowledge and deliberate reasoning. Contains 11.5K meticulously collected multimodal questions from college exams, quizzes, and textbooks, covering six core disciplines: Art & Design, Business, Science, Health & Medicine, Humanities & Social Science, and Tech & Engineering across 30 subjects and 183 subfields.
Qwen3.6 Plus from Alibaba Cloud / Qwen Team currently leads the MMMU leaderboard with a score of 0.860 across 62 evaluated AI models.
Qwen3.6 Plus leads with 86.0%, followed by 
GPT-5.1 at 85.4% and GPT-5.1 Thinking at 85.4%.
Progress Over Time
Interactive timeline showing model performance evolution on MMMU
MMMU Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 |  Alibaba Cloud / Qwen Team | — | 1.0M | $0.50 / $3.00 | ||
| 2 | OpenAI | — | 400K | $1.25 / $10.00 | ||
| 2 | OpenAI | — | 400K | $1.25 / $10.00 | ||
| 2 | OpenAI | — | 400K | $1.25 / $10.00 | ||
| 5 | OpenAI | — | — | — | ||
| 6 | Alibaba Cloud / Qwen Team | 122B | 262K | $0.40 / $3.20 | ||
| 7 | OpenAI | — | — | — | ||
| 7 | Alibaba Cloud / Qwen Team | 28B | 262K | $0.60 / $3.60 | ||
| 9 | Alibaba Cloud / Qwen Team | 27B | 262K | $0.30 / $2.40 | ||
| 10 | — | — | — | |||
| 11 | Alibaba Cloud / Qwen Team | 35B | — | — | ||
| 12 | OpenAI | — | — | — | ||
| 13 | Alibaba Cloud / Qwen Team | 35B | 262K | $0.25 / $2.00 | ||
| 14 |  Google | — | 1.0M | $0.30 / $2.50 | ||
| 15 | Google | — | 1.0M | $1.25 / $10.00 | ||
| 16 |  StepFun | 10B | — | — | ||
| 17 |  xAI | — | 128K | $3.00 / $15.00 | ||
| 18 | OpenAI | — | — | — | ||
| 19 | — | — | — | |||
| 20 | OpenAI | — | — | — | ||
| 21 |  Anthropic | — | — | — | ||
| 22 | OpenAI | — | 1.0M | $2.00 / $8.00 | ||
| 23 | Anthropic | — | — | — | ||
| 24 |  Meta | 400B | — | — | ||
| 25 | Google | — | — | — | ||
| 26 | OpenAI | — | 1.0M | $0.40 / $1.60 | ||
| 27 | OpenAI | — | 128K | $2.50 / $10.00 | ||
| 28 | Google | — | — | — | ||
| 29 | Alibaba Cloud / Qwen Team | 73B | — | — | ||
| 30 | Alibaba Cloud / Qwen Team | 72B | — | — | ||
| 31 | Alibaba Cloud / Qwen Team | 34B | — | — | ||
| 31 |  Moonshot AI | — | — | — | ||
| 33 | Meta | 109B | — | — | ||
| 34 | Anthropic | — | — | — | ||
| 35 | Google | — | — | — | ||
| 36 | xAI | — | — | — | ||
| 37 | Google | — | — | — | ||
| 38 |  Mistral AI | 124B | — | — | ||
| 39 | xAI | — | — | — | ||
| 40 | Mistral AI | 24B | — | — | ||
| 41 | Google | — | — | — | ||
| 42 |  Amazon | — | — | — | ||
| 43 | 90B | — | — | |||
| 44 | OpenAI | — | — | — | ||
| 45 | Mistral AI | 24B | — | — | ||
| 45 | Mistral AI | 24B | — | — | ||
| 47 | Alibaba Cloud / Qwen Team | 7B | — | — | ||
| 48 | Alibaba Cloud / Qwen Team | 8B | — | — | ||
| 49 | Amazon | — | — | — | ||
| 50 | OpenAI | — | 1.0M | $0.10 / $0.40 |
FAQ
Common questions about MMMU.
More evaluations to explore
Related benchmarks in the same category
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks. Features over 12,000 curated questions across 14 domains and causes a 16-33% accuracy drop compared to original MMLU.
All 30 problems from the 2025 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999. Used as an AI benchmark to evaluate large language models' ability to solve complex mathematical problems requiring multi-step logical deductions and structured symbolic reasoning.
Massive Multitask Language Understanding benchmark testing knowledge across 57 diverse subjects including STEM, humanities, social sciences, and professional domains
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions