Evals for Every Language - Language arz: leaderboard
Metric: Average Score (%). Source: huggingface.co. 71 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 54.92 |
| 2 | Claude Opus 4.7 | 52.06 |
| 3 | Gemini 2.5 Pro | 51.94 |
| 4 | Claude Sonnet 4.5 | 51.32 |
| 5 | Gemini 3.1 Flash Lite | 49.85 |
| 6 | Claude Opus 4.8 | 49.71 |
| 7 | Gemini 2.5 Flash | 48.42 |
| 8 | GPT-4o | 48.32 |
| 9 | GPT-5.3 | 48.26 |
| 10 | Claude Sonnet 4 | 48.06 |
| 11 | GPT-5.1 | 47.84 |
| 12 | GPT-5 | 47.48 |
| 13 | Command A | 47.15 |
| 14 | Qwen 3.6 Plus | 46.8 |
| 15 | Nova Pro (v1) | 46.63 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-arz · How It Works · Data refreshed daily, snapshot 2026-09-05.