Evals for Every Language - Language ab: leaderboard
Metric: Average Score (%). Source: huggingface.co. 67 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 100 |
| 2 | GPT-5.5 | 100 |
| 3 | Claude Opus 4.8 | 100 |
| 4 | GPT-5.4 | 96.67 |
| 5 | Claude Opus 4.7 | 96.67 |
| 6 | GPT-5.3 | 93.33 |
| 7 | GPT-5.2 | 90 |
| 8 | GPT-5 | 86.67 |
| 9 | Gemini 3.1 Flash Lite | 82.22 |
| 10 | Claude Sonnet 4.6 | 81.48 |
| 11 | GPT-4.1 | 80 |
| 12 | Grok 4.20 | 80 |
| 13 | DeepSeek V3.1 | 75 |
| 14 | GPT-5.1 | 73.33 |
| 15 | nova-2-lite-v1 | 66.67 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-ab · How It Works · Data refreshed daily, snapshot 2026-09-05.