📚 Knowledge · llm-stats.com
Multilingual MMLU
MMLU translated across 14 languages.
12
Modelli testati
Claude Mythos Preview
Modello di punta
92.7%
Punteggio migliore
Anthropic
Provider
| # | Modello | Multilingual MMLU |
|---|---|---|
| 1 | Claude Mythos Preview Anthropic | 92.7% |
| 2 | Gemini 3.1 Pro Google | 92.6% |
| 3 | Gemini 3 Pro Google | 91.8% |
| 4 | Gemini 3 Flash Google | 91.8% |
| 5 | Claude Opus 4.7 Anthropic | 91.5% |
| 6 | Claude Opus 4.6 Anthropic | 91.1% |
| 7 | Qwen3.7 Max Alibaba | 90.3% |
| 8 | GPT-5.2 OpenAI | 89.6% |
| 9 | Qwen3.6 Plus Alibaba | 89.5% |
| 10 | Claude Sonnet 4.6 Anthropic | 89.3% |
| 11 | Qwen3.7 Plus Alibaba | 89% |
| 12 | Qwen3.5 397B-A17B Alibaba | 88.5% |
Dati benchmark aggregati da llm-stats.com. Vedi Metodologia.