Evals for Every Language - Language new: leaderboard
Metric: Average Score (%). Source: huggingface.co. 67 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 96.3 |
| 2 | Qwen 3.6 Plus | 93.33 |
| 3 | Gemini 2.5 Pro | 90 |
| 4 | Claude Opus 4.7 | 90 |
| 5 | Claude Opus 4.8 | 90 |
| 6 | Gemini 3.1 Flash Lite | 88.89 |
| 7 | Claude Sonnet 4.6 | 87.5 |
| 8 | GPT-5.5 | 86.67 |
| 9 | Qwen 3.6 Flash | 86.67 |
| 10 | Grok 4.20 | 80 |
| 11 | GPT-5.2 | 73.33 |
| 12 | Gemini 2.5 Flash Lite | 63.33 |
| 13 | granite-4.0-h-micro | 62.5 |
| 14 | GPT-5.4 | 61.11 |
| 15 | Qwen 3 30B A3B 2507 Instruct | 58.33 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-new · How It Works · Data refreshed daily, snapshot 2026-09-05.