Evals for Every Language - Cantonese (yue): leaderboard
Average score across evaluation tasks in Cantonese (yue), from Evals for Every Language.
Metric: Average Score (%). Source: huggingface.co. Saturation forecast: Around 2030. 65 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 | 71.78 | #37 |
| 2 | Claude Opus 4.7 | 71.59 | #49 |
| 3 | Claude Sonnet 4.6 | 64.36 | #90 |
| 4 | Gemini 3.1 Pro (Preview) | 59.57 | #58 |
| 5 | Gemini 2.5 Pro | 57.98 | #152 |
| 6 | Claude Sonnet 4.5 | 57.95 | #142 |
| 7 | Claude Sonnet 4 | 57.55 | #199 |
| 8 | GPT-5.1 | 56.66 | #135 |
| 9 | DeepSeek V3.2 Exp | 55.9 | #234 |
| 10 | GLM-4 32B (0414) | 55.77 | #533 |
| 11 | GPT-5.2 | 55.73 | #110 |
| 12 | Kimi K2 | 55.73 | #243 |
| 13 | GPT-5.3 | 55.73 | #103 |
| 14 | Gemini 2.5 Flash Lite | 55.66 | #422 |
| 15 | Gemma 3 27B (IT) | 55.5 | #523 |
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-yue · How It Works · Data refreshed daily, snapshot 2026-10-20.