Evals for Every Language - Language mai: leaderboard
Metric: Average Score (%). Source: huggingface.co. 71 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 70.54 |
| 2 | Gemini 2.5 Pro | 69.95 |
| 3 | Claude Sonnet 4.5 | 69.23 |
| 4 | GPT-5.3 | 69.03 |
| 5 | Claude Sonnet 4 | 68.61 |
| 6 | Claude Opus 4.7 | 68.5 |
| 7 | Claude Opus 4.8 | 68.44 |
| 8 | GPT-5.1 | 68.44 |
| 9 | GPT-5.5 | 68.16 |
| 10 | GPT-5 | 67.81 |
| 11 | GPT-5.2 | 66.76 |
| 12 | Qwen 3.6 Plus | 66.65 |
| 13 | GPT-4o | 66.29 |
| 14 | Grok 4.20 | 66.2 |
| 15 | GPT-4.1 | 65.99 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-mai · How It Works · Data refreshed daily, snapshot 2026-09-05.