llmboard.aiLeaderboard Center
Overall
Overall RankingOpen Models
Tools
Model DirectoryCompare Models
Capabilities
CodingReasoningMathKnowledgeInstruction Following
Price & Efficiency
Price & ValueCapability vs. PriceRuntime Performance
Modalities
Image GenerationVideo GenerationSpeech ModelsEmbeddings
Core Benchmarks
GPQAMMLU-ProAIME 2025SWE-Bench VerifiedMMLUHumanity's Last ExamLiveCodeBenchMATHHumanEvalMMMU-ProView all benchmarks
Methods
Scoring & Data
393 models668 benchmarks

Leaderboard Center

Overall RankingCodingCore BenchmarksPrice & ValueRuntime Performance

Modalities

All ModelsImage GenerationImage EditingVideo GenerationImage-to-VideoVideo EditingText-to-SpeechSpeech-to-TextEmbeddings

Data & Methods

Scoring MethodAll BenchmarksReasoningMath

Vendors

All VendorsOpenAIAnthropicGoogle
llmboard.aiCopyright 2026 llmboard.ai

language benchmark

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.

Updated Aug 11, 2026

Models49
Model coverage49
MetricScore
EvidenceB

On this page

  • Ranking
  • Distribution
  • Highlights
  • About
  • FAQ

MMMLU Ranking

Higher score ranks better on this benchmark.

49 rows
Columns

Show columns

01ANClaude Mythos PreviewAnthropic92.7%100.0%49CAug 11, 2026
02GOGemini 3.1 ProGoogle92.6%97.9%49CAug 11, 2026
03GOGemini 3 FlashGoogle91.8%95.8%49CAug 11, 2026
04GOGemini 3 ProGoogle91.8%93.8%49CAug 11, 2026
05ANClaude Opus 4.7Anthropic91.5%91.7%49CAug 11, 2026
06ANClaude Opus 4.6Anthropic91.1%89.6%49CAug 11, 2026
07ANClaude Opus 4.5Anthropic90.8%87.5%49CAug 11, 2026
08ACQwen3.7 MaxAlibaba Cloud / Qwen Team90.3%85.4%49CAug 11, 2026
09OPGPT-5.2OpenAI89.6%83.3%49CAug 11, 2026
10ANClaude Opus 4.1Anthropic89.5%81.3%49CAug 11, 2026
11ACQwen3.6 PlusAlibaba Cloud / Qwen Team89.5%79.2%49CAug 11, 2026
12ANClaude Sonnet 4.6Anthropic89.3%77.1%49CAug 11, 2026
13ANClaude Sonnet 4.5Anthropic89.1%75.0%49CAug 11, 2026
14ACQwen3.7-PlusAlibaba Cloud / Qwen Team89.0%72.9%49CAug 11, 2026
15GOGemini 3.1 Flash-LiteGoogle88.9%70.8%49CAug 11, 2026
16ANClaude Opus 4Anthropic88.8%68.8%49CAug 11, 2026
17ACQwen3.5-397B-A17BAlibaba Cloud / Qwen Team88.5%66.7%49CAug 11, 2026
18GOGemma 4 31BGoogle88.4%64.6%49CAug 11, 2026
19OPo1OpenAI87.7%62.5%49CAug 11, 2026
20OPGPT-4.1OpenAI87.3%60.4%49CAug 11, 2026
21ACQwen3 235B A22BAlibaba Cloud / Qwen Team86.7%58.3%49CAug 11, 2026
22ACQwen3.5-122B-A10BAlibaba Cloud / Qwen Team86.7%56.3%49CAug 11, 2026
23ANClaude Sonnet 4Anthropic86.5%54.2%49CAug 11, 2026
24GOGemma 4 26B-A4BGoogle86.3%52.1%49CAug 11, 2026
25ANClaude 3.7 SonnetAnthropic86.1%50.0%49CAug 11, 2026
26ACQwen3.5-27BAlibaba Cloud / Qwen Team85.9%47.9%49CAug 11, 2026
27LAK-EXAONE-236B-A23BLG AI Research85.7%45.8%49CAug 11, 2026
28MAMistral Large 3 (675B Instruct 2512 Eagle)Mistral AI85.5%43.8%49CAug 11, 2026
29MAMistral Large 3 (675B Base)Mistral AI85.5%41.7%49CAug 11, 2026
30MAMistral Large 3 (675B Instruct 2512 NVFP4)Mistral AI85.5%39.6%49CAug 11, 2026
31MAMistral Large 3 (675B Instruct 2512)Mistral AI85.5%37.5%49CAug 11, 2026
32ACQwen3.5-35B-A3BAlibaba Cloud / Qwen Team85.2%35.4%49CAug 11, 2026
33OPGPT-4.5OpenAI85.1%33.3%49CAug 11, 2026
34OPGPT OSS 120B HighOpenAI83.8%31.3%49CAug 11, 2026
35GOGemma 4 12BGoogle83.4%29.2%49CAug 11, 2026
36ANClaude Haiku 4.5Anthropic83.0%27.1%49CAug 11, 2026
37GODiffusionGemma 26B-A4BGoogle81.5%25.0%49CAug 11, 2026
38OPGPT-4oOpenAI81.4%22.9%49CAug 11, 2026
39ACQwen3.5-9BAlibaba Cloud / Qwen Team81.2%20.8%49CAug 11, 2026
40OPGPT-4.1 miniOpenAI78.5%18.8%49CAug 11, 2026
41GOGemma 4 E4BGoogle76.6%16.7%49CAug 11, 2026
42ACQwen3.5-4BAlibaba Cloud / Qwen Team76.1%14.6%49CAug 11, 2026
43MAMistral Large 3Mistral AI74.2%12.5%49CAug 11, 2026
44MIPhi-3.5-MoE-instructMicrosoft69.9%10.4%49CAug 11, 2026
45GOGemma 4 E2BGoogle67.4%8.3%49CAug 11, 2026
46OPGPT-4.1 nanoOpenAI66.9%6.3%49CAug 11, 2026
47ACQwen3.5-2BAlibaba Cloud / Qwen Team63.1%4.2%49CAug 11, 2026
48MIPhi-3.5-mini-instructMicrosoft55.4%2.1%49CAug 11, 2026
49ACQwen3.5-0.8BAlibaba Cloud / Qwen Team44.3%0.0%49CAug 11, 2026

MMMLU Score Distribution

A closer view of the leading scores on this benchmark.

MMMLU

MMMLU Highlights

The leading models and scores on this benchmark.

Rank #1Claude Mythos Preview92.7%Rank #2Gemini 3.1 Pro92.6%Rank #3Gemini 3 Flash91.8%Rank #4Gemini 3 Pro91.8%

What is MMMLU?

What MMMLU measures and how its scores work.

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.

Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.

Family
MMMLU
Modality
text
Primary category
language
Score direction
higher
LLMBoard eligible
Yes
Evaluation key
mmmlu|llm-stats-current

Benchmark scores retain their original unit. Overall score eligibility is shown separately.

FAQ

Common questions about MMMLU.

Which model scores highest on MMMLU?

Claude Mythos Preview is currently ranked first with 92.7%.

What does MMMLU measure?

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.

Is a higher score better?

Yes. Higher values rank better for this benchmark.

How many models are compared?

49 model results are currently shown.

Does this benchmark affect the overall score?

Yes. This benchmark can contribute to the current LLMBoard capability score.