Evals for Every Language - Language tum: leaderboard
Metric: Average Score (%). Source: huggingface.co. 70 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 49.09 |
| 2 | Gemini 3.1 Pro (Preview) | 49.01 |
| 3 | Gemini 3.1 Flash Lite | 48.15 |
| 4 | Gemini 2.5 Flash | 47.53 |
| 5 | Claude Opus 4.7 | 46.54 |
| 6 | GPT-5.5 | 46.28 |
| 7 | GPT-5.3 | 44.53 |
| 8 | GPT-5.4 | 43.09 |
| 9 | GPT-4o | 42.19 |
| 10 | Claude Sonnet 4.5 | 41.92 |
| 11 | GPT-5.1 | 41.17 |
| 12 | Claude Sonnet 4 | 40.9 |
| 13 | Claude Opus 4.8 | 39.86 |
| 14 | GPT-5 | 39.61 |
| 15 | Nova Pro (v1) | 39.08 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-tum · How It Works · Data refreshed daily, snapshot 2026-09-05.