Evals for Every Language - Language ckb: leaderboard
Metric: Average Score (%). Source: huggingface.co. 71 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 74.15 |
| 2 | Claude Opus 4.8 | 71.59 |
| 3 | Claude Opus 4.7 | 70.88 |
| 4 | Gemini 2.5 Pro | 70 |
| 5 | GPT-5.5 | 68.48 |
| 6 | Qwen 3.6 Plus | 68.26 |
| 7 | Gemini 3.1 Flash Lite | 67.79 |
| 8 | GPT-5.3 | 67.73 |
| 9 | Claude Sonnet 4 | 65.2 |
| 10 | Claude Sonnet 4.5 | 65.15 |
| 11 | DeepSeek V3.1 | 64.65 |
| 12 | Llama 4 Maverick | 64.58 |
| 13 | Claude Sonnet 4.6 | 64.34 |
| 14 | Grok 4.20 | 64.15 |
| 15 | DeepSeek V3.2 Exp | 63.03 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-ckb · How It Works · Data refreshed daily, snapshot 2026-09-05.