AI benchmarks and leaderboards are different things
Learn the difference between a benchmark, an evaluation run and a leaderboard, and use each to answer the right question.
Model reports
MMMLU
A language-count average. Two cards reporting "MMMLU" may be averaging over different language sets, so the counts must match to compare.
Evaluation context and evidence
- Source
- Model reports
- Publisher
- Not recorded
- Benchmark release
- 2024-09-24
- Collected
- Not recorded
- Modality
- Not recorded
These are source-reported observations, not independent reruns by this site. Check the task version, evaluation protocol, unit and reporting date before comparing models. Results from different setups are not a universal ranking.
Original sources
Reported results 5
Shown in source observation order, not ranked. Dates retain the precision provided by the source.
| Model | Value | Unit | Protocol | Reported date and basis | Source |
|---|---|---|---|---|---|
| Claude Opus 4 | 88.8 | percent | 14 non-English languages, extended thinking up to 64K tokens | 2025-05-22 · document_publication | Claude Opus 4 and Sonnet 4 |
| Claude Sonnet 4 | 86.5 | percent | 14 non-English languages, extended thinking up to 64K tokens | 2025-05-22 · document_publication | Claude Opus 4 and Sonnet 4 |
| Gemini 3.1 Pro | 92.6 | percent | Thinking (High) | 2026-02-19 · document_publication | Gemini 3.1 Pro |
| Gemma 4 (31B) | 88.4 | percent | as reported | 2026-03-11 · document_publication | Gemma 4 (31B) |
| Qwen3.5-397B-A17B | 88.5 | percent | thinking (temp 0.6, top_p 0.95, top_k 20) | 2026-02-16 · document_publication | Qwen3.5-397B-A17B |