Cedar - Multimodal - English: leaderboard
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around 2030. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 VL 235B A22B (Thinking) | 50.38 |
| 2 | Gemini 2.5 Flash (Thinking) | 50.28 |
| 3 | Claude Sonnet 4.5 | 49.62 |
| 4 | Gemma 3 27B (IT) | 48.37 |
| 5 | Mistral Small 3.2 | 47.85 |
| 6 | aya-vision-8B | 42.37 |
| 7 | Qwen 3 VL 8B (Thinking) | 41.88 |
| 8 | Pixtral-12B-2409 | 38.01 |
| 9 | GPT-4o (2024-11-20) | 36.34 |
| 10 | Llama 3.2 11B Instruct | 34.29 |
Interactive version: theaggregate.ai/benchmark?slug=cedar-multimodal-english · How It Works · Data refreshed daily, snapshot 2026-09-25.