Evals for Every Language - Language ary: leaderboard
Metric: Average Score (%). Source: huggingface.co. 71 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 48.93 |
| 2 | GPT-5.5 | 47.34 |
| 3 | Gemini 2.5 Flash Lite | 47.32 |
| 4 | Gemini 2.5 Pro | 46.95 |
| 5 | Claude Sonnet 4.5 | 46.35 |
| 6 | GPT-5.3 | 46.15 |
| 7 | GPT-5.2 | 45.47 |
| 8 | Claude Sonnet 4 | 45.41 |
| 9 | GPT-4o | 44.98 |
| 10 | GPT-5.1 | 44.92 |
| 11 | Gemini 2.5 Flash | 44.39 |
| 12 | Claude Opus 4.7 | 44.23 |
| 13 | Qwen 3.6 Plus | 42.5 |
| 14 | Claude Sonnet 4.6 | 42.37 |
| 15 | Gemma 3 27B (IT) | 42.15 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-ary · How It Works · Data refreshed daily, snapshot 2026-09-05.