Evals for Every Language - Language wo: leaderboard
Metric: Average Score (%). Source: huggingface.co. 70 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 56.45 |
| 2 | Claude Opus 4.8 | 52.76 |
| 3 | GPT-5.2 | 50.89 |
| 4 | Claude Opus 4.7 | 50.76 |
| 5 | GPT-5 | 49.71 |
| 6 | GPT-5.5 | 48.68 |
| 7 | Gemini 3.1 Flash Lite | 46.31 |
| 8 | Gemini 2.5 Pro | 46.09 |
| 9 | Claude Sonnet 4.6 | 45.26 |
| 10 | DeepSeek V3.1 | 44.91 |
| 11 | GPT-5.4 | 43.84 |
| 12 | GPT-5.3 | 43.57 |
| 13 | GPT-5.1 | 42.64 |
| 14 | Qwen 3.6 Plus | 40.55 |
| 15 | GPT-4o | 40.32 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-wo · How It Works · Data refreshed daily, snapshot 2026-09-05.