Evals for Every Language - Language mos: leaderboard
Metric: Average Score (%). Source: huggingface.co. 70 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 40.3 |
| 2 | Claude Opus 4.7 | 35.22 |
| 3 | Claude Sonnet 4.5 | 35.06 |
| 4 | Gemini 3.1 Flash Lite | 34.94 |
| 5 | Claude Sonnet 4.6 | 34.91 |
| 6 | Gemini 2.5 Pro | 34.25 |
| 7 | Grok 4.20 | 32.6 |
| 8 | GPT-5.3 | 31.83 |
| 9 | GPT-5.5 | 31.1 |
| 10 | Qwen 3.6 Plus | 30.65 |
| 11 | GPT-5.4 | 30.29 |
| 12 | Claude Opus 4.8 | 28.74 |
| 13 | GPT-4o | 28.03 |
| 14 | Claude Sonnet 4 | 27.21 |
| 15 | GPT-5.1 | 27.1 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-mos · How It Works · Data refreshed daily, snapshot 2026-09-05.