| Benchmark | Model | Rank | Results |
|---|---|---|---|
| video-question-answering-on-mlvu | CAt+v (Qwen2.5-VL, 3B) | #42 | Accuracy: 67.9 |
| optical-character-recognition-ocr-on-ocrbench | CAQ2.5-VL | #55 | Score: 783.0 |
| chart-question-answering-on-chartqa | CAt+v (Qwen2.5-VL, 3B) | #64 | 1:1 Accuracy: 81.3 |
| video-question-answering-on-mvbench | CAt+v (Qwen2.5-VL, 3B) | #65 | Avg.: 60.6 |
| visual-question-answering-vqa-on-realworldqa | CAt+v (Qwen2.5-VL, 3B) | #74 | Accuracy: 63.8 |