Evals for Every Language - Mari (chm): leaderboard
Average score across evaluation tasks in Mari (chm), from Evals for Every Language.
Metric: Average Score (%). Source: huggingface.co. Saturation forecast: Around January 2027. 64 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 | 69.72 | #37 |
| 2 | Claude Opus 4.7 | 69.52 | #49 |
| 3 | Gemini 3.1 Flash Lite | 67.2 | #191 |
| 4 | Gemini 3.1 Pro (Preview) | 66 | #58 |
| 5 | Claude Sonnet 4.6 | 65.38 | #90 |
| 6 | GPT-5.4 | 64.25 | #79 |
| 7 | Qwen 3.6 Plus | 63.93 | #120 |
| 8 | GPT-5.5 | 63.03 | #31 |
| 9 | Qwen 3.6 Flash | 61.66 | #214 |
| 10 | Grok 4.20 | 61.04 | #175 |
| 11 | GPT-5.3 | 60.98 | #103 |
| 12 | Gemini 2.5 Pro | 60.24 | #152 |
| 13 | Claude Sonnet 4.5 | 59.41 | #142 |
| 14 | GPT-5.2 | 53.86 | #110 |
| 15 | Claude Sonnet 4 | 53.58 | #199 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-chm · How It Works · Data refreshed daily, snapshot 2026-10-20.