Evals for Every Language - Language so: leaderboard
Metric: Average Score (%). Source: huggingface.co. 71 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 | 69.48 |
| 2 | Claude Opus 4.8 | 69.43 |
| 3 | Gemini 3.1 Pro (Preview) | 69.39 |
| 4 | Claude Sonnet 4.5 | 68.12 |
| 5 | GPT-4o | 67.55 |
| 6 | GPT-5.2 | 66.8 |
| 7 | GPT-5.3 | 66.66 |
| 8 | Grok 4.20 | 66.6 |
| 9 | Claude Sonnet 4.6 | 66.54 |
| 10 | GPT-5.1 | 66.14 |
| 11 | GPT-5 | 65.71 |
| 12 | Gemini 3.1 Flash Lite | 65.44 |
| 13 | Qwen 3.6 Plus | 65.41 |
| 14 | GPT-5.5 | 65.36 |
| 15 | GPT-4.1 | 65.28 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-so · How It Works · Data refreshed daily, snapshot 2026-09-05.