Cedar - Text-Only - Chinese: leaderboard
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around 2030. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.5 | 47.51 |
| 2 | Qwen 3 VL 235B A22B (Thinking) | 43.22 |
| 3 | Gemini 2.5 Flash (Thinking) | 43.12 |
| 4 | Gemma 3 27B (IT) | 40.85 |
| 5 | Kimi K2 | 40.8 |
| 6 | GPT-4o (2024-11-20) | 40.34 |
| 7 | aya-vision-8B | 37.43 |
| 8 | Qwen 3 VL 8B (Thinking) | 33.9 |
| 9 | Mistral Small 3.2 | 32.96 |
| 10 | Pixtral-12B-2409 | 31.81 |
Interactive version: theaggregate.ai/benchmark?slug=cedar-text-only-chinese · How It Works · Data refreshed daily, snapshot 2026-09-25.