Evals for Every Language - Language li: leaderboard
Metric: Average Score (%). Source: huggingface.co. 71 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 73.08 |
| 2 | Claude Sonnet 4.5 | 72.44 |
| 3 | Gemini 3.1 Flash Lite | 72.06 |
| 4 | Gemini 2.5 Pro | 71.66 |
| 5 | Claude Sonnet 4.6 | 71.08 |
| 6 | Claude Opus 4.8 | 70.76 |
| 7 | GPT-5 | 70.32 |
| 8 | GPT-5.3 | 70.24 |
| 9 | Qwen 3.6 Plus | 70.22 |
| 10 | GPT-5.2 | 70.16 |
| 11 | Claude Sonnet 4 | 69.58 |
| 12 | GPT-5.4 | 69.13 |
| 13 | GPT-5.5 | 69.1 |
| 14 | Nova Pro (v1) | 68.94 |
| 15 | Gemma 3 27B (IT) | 68.87 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-li · How It Works · Data refreshed daily, snapshot 2026-09-05.