Evals for Every Language - Language pam — leaderboard
Metric: Average Score (%). Source: huggingface.co. 68 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 96.67 |
| 2 | GPT-5 | 96.67 |
| 3 | Claude Opus 4.7 | 96.67 |
| 4 | Claude Opus 4.8 | 96.67 |
| 5 | GPT-5.1 | 96.67 |
| 6 | Grok 4.20 | 96.67 |
| 7 | DeepSeek V3.1 | 94.44 |
| 8 | Claude Sonnet 4.5 | 93.33 |
| 9 | GPT-5.5 | 93.33 |
| 10 | GPT-5.4 | 93.33 |
| 11 | Qwen 3.6 Plus | 93.33 |
| 12 | Claude Sonnet 4.6 | 92.96 |
| 13 | GPT-4o | 90 |
| 14 | DeepSeek V3.2 Exp | 87.5 |
| 15 | GPT-5.2 | 86.67 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-pam · How the rankings work · Data refreshed daily, snapshot 2026-07-22.