Evals for Every Language - Language is: leaderboard
Metric: Average Score (%). Source: huggingface.co. 71 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Flash Lite | 73.86 |
| 2 | Gemini 3.1 Pro (Preview) | 72.85 |
| 3 | Gemini 2.5 Pro | 72.39 |
| 4 | Claude Sonnet 4 | 72.37 |
| 5 | Claude Sonnet 4.5 | 72.23 |
| 6 | GPT-5 | 72.12 |
| 7 | Claude Opus 4.7 | 72.09 |
| 8 | Claude Haiku 4.5 | 70.92 |
| 9 | GPT-5.1 | 70.85 |
| 10 | Kimi K2 | 70.59 |
| 11 | GPT-5.3 | 70.54 |
| 12 | GPT-4o | 70.42 |
| 13 | Claude Opus 4.8 | 70.35 |
| 14 | Grok 4.20 | 70.07 |
| 15 | GPT-4.1 | 69.6 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-is · How It Works · Data refreshed daily, snapshot 2026-09-05.