Cedar - Text-Only - Japanese: leaderboard
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around 2033. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o (2024-11-20) | 42.97 |
| 2 | Kimi K2 | 39.66 |
| 3 | aya-vision-8B | 37.8 |
| 4 | Qwen 3 VL 235B A22B (Thinking) | 37.19 |
| 5 | Gemini 2.5 Flash (Thinking) | 36.31 |
| 6 | Gemma 3 27B (IT) | 36.03 |
| 7 | Pixtral-12B-2409 | 33.16 |
| 8 | Mistral Small 3.2 | 31.77 |
| 9 | Claude Sonnet 4.5 | 29.6 |
| 10 | Qwen 3 VL 8B (Thinking) | 28.24 |
Interactive version: theaggregate.ai/benchmark?slug=cedar-text-only-japanese · How It Works · Data refreshed daily, snapshot 2026-09-25.