MMMLU
Multilingual Massive Multitask Language Understanding dataset released by OpenAI, featuring professionally translated MMLU test questions across 14 languages including Arabic, Bengali, German, Spanish, French, Hindi, Indonesian, Italian, Japanese, Korean, Portuguese, Swahili, Yoruba, and Chinese. Contains approximately 15,908 multiple-choice questions per language covering 57 subjects.
Claude Mythos Preview from Anthropic currently leads the MMMLU leaderboard with a score of 0.927 across 46 evaluated AI models.
What MMMLU measures
MMMLU is a text benchmark that evaluates large language models on general, language, math, and reasoning tasks. LLM Stats tracks 46 models on this benchmark, with a maximum possible score of 1. Current average across reported models is 0.8, with the leader reaching 0.9.
Compare leaders on the best AI for general, best AI for language, best AI for math and best AI for reasoning leaderboards.
Publication
- Paper
- Measuring Massive Multitask Language Understanding
- Authors
- Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, and 3 others
- Published
- arXiv
- 2009.03300
Abstract
We propose a new test to measure a text model's multitask accuracy. The test covers 57 tasks including elementary mathematics, US history, computer science, law, and more. To attain high accuracy on this test, models must possess extensive world knowledge and problem solving ability. We find that while most recent models have near random-chance accuracy, the very largest GPT-3 model improves over random chance by almost 20 percentage points on average. However, on every one of the 57 tasks, the best models still need substantial improvements before they can reach expert-level accuracy. Models also have lopsided performance and frequently do not know when they are wrong. Worse, they still have near-random accuracy on some socially important subjects such as morality and law. By comprehensively evaluating the breadth and depth of a model's academic and professional understanding, our test can be used to analyze models across many tasks and to identify important shortcomings.
Claude Mythos Preview leads with 92.7%, followed by 
Gemini 3.1 Pro at 92.6% and Gemini 3 Pro at 91.8%.
Progress Over Time
Interactive timeline showing model performance evolution on MMMLU
MMMLU Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 | Anthropic | — | — | — | ||
| 2 | Google | — | 1.0M | $2.50 / $15.00 | ||
| 3 | Google | — | — | — | ||
| 3 | Google | — | 1.0M | $0.50 / $3.00 | ||
| 5 | Anthropic | — | 1.0M | $5.00 / $25.00 | ||
| 6 | Anthropic | — | 1.0M | $5.00 / $25.00 | ||
| 7 | Anthropic | — | — | — | ||
| 8 |  Alibaba Cloud / Qwen Team | — | 1.0M | $1.25 / $3.75 | ||
| 9 |  OpenAI | — | 400K | $1.75 / $14.00 | ||
| 10 | Alibaba Cloud / Qwen Team | — | 1.0M | $0.50 / $3.00 | ||
| 10 | Anthropic | — | — | — | ||
| 12 | Anthropic | — | 200K | $3.00 / $15.00 | ||
| 13 | Anthropic | — | 200K | $3.00 / $15.00 | ||
| 14 | Google | — | 1.0M | $0.25 / $1.50 | ||
| 15 | Anthropic | — | — | — | ||
| 16 | Alibaba Cloud / Qwen Team | 397B | 262K | $0.60 / $3.60 | ||
| 17 | Google | 31B | 262K | $0.14 / $0.40 | ||
| 18 | OpenAI | — | — | — | ||
| 19 | OpenAI | — | 1.0M | $2.00 / $8.00 | ||
| 20 | Alibaba Cloud / Qwen Team | 235B | — | — | ||
| 20 | Alibaba Cloud / Qwen Team | 122B | 262K | $0.40 / $3.20 | ||
| 22 | Anthropic | — | — | — | ||
| 23 | Google | 25B | 262K | $0.13 / $0.40 | ||
| 24 | Anthropic | — | — | — | ||
| 25 | Alibaba Cloud / Qwen Team | 27B | 262K | $0.30 / $2.40 | ||
| 26 |  LG AI Research | 236B | — | — | ||
| 27 |  Mistral AI | 675B | 262K | $0.50 / $1.50 | ||
| 27 | 675B | — | — | |||
| 27 | Mistral AI | 675B | — | — | ||
| 27 | 675B | — | — | |||
| 31 | Alibaba Cloud / Qwen Team | 35B | 262K | $0.25 / $2.00 | ||
| 32 | OpenAI | — | — | — | ||
| 33 | OpenAI | 117B | 131K | $0.10 / $0.50 | ||
| 34 | Anthropic | — | 200K | $1.00 / $5.00 | ||
| 35 | OpenAI | — | 128K | $2.50 / $10.00 | ||
| 36 | Alibaba Cloud / Qwen Team | 9B | — | — | ||
| 37 | OpenAI | — | 1.0M | $0.40 / $1.60 | ||
| 38 | Google | 8B | — | — | ||
| 39 | Alibaba Cloud / Qwen Team | 4B | — | — | ||
| 40 | Mistral AI | 675B | — | — | ||
| 41 |  Microsoft | 60B | — | — | ||
| 42 | Google | 5B | — | — | ||
| 43 | OpenAI | — | 1.0M | $0.10 / $0.40 | ||
| 44 | Alibaba Cloud / Qwen Team | 2B | — | — | ||
| 45 | Microsoft | 4B | — | — | ||
| 46 | Alibaba Cloud / Qwen Team | 800M | — | — |
FAQ
Common questions about MMMLU.
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