Chinese Culture Multimodal Evaluation on CCBench (dev)
71.2AccuracyStep-1V
Evaluation results
| Method | Links | |
|---|---|---|
| Step-1V#param=100B, open-source=false2024.04 | 71.2 | |
| InternVL 1.5#param=26B, open-source=false2024.04 | 69.8 | |
| Qwen-VL-Maxopen-source=false2024.04 | 63.5 | |
| InternVL 1.2#param=40B, open-source=false2024.04 | 59.2 | |
| Qwen-VL-Plusopen-source=false2024.04 | 55.1 | |
| Gemini Pro 1.0open-source=false2024.04 | 52.5 | |
| LLaVA-NeXT#param=35B, open-source=false2024.04 | 49.2 | |
| MM1#param=30B, open-source=false2024.04 | 48.7 | |
| GPT-4Vopen-source=false2024.04 | 46.5 | |
| Claude-3 Sonnetopen-source=false2024.04 | 27.8 | |
| Claude-3 Opusopen-source=false2024.04 | 26.3 | |
| Claude-3 Haikuopen-source=false2024.04 | 24.5 |