| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Video Understanding | MotionBench (val) | Accuracy65.4 | 21 | |
| Motion-level Perception | MotionBench (test) | Accuracy58 | 14 | |
| Motion-level Perception | MotionBench (dev) | Accuracy58 | 14 | |
| Motion Understanding | MotionBench | Accuracy61.4 | 10 | |
| Video Question Answering | MotionBench | ALL47.41 | 3 | |
| Short video understanding | MotionBench | Accuracy68.4 | 3 | |
| Video Captioning | MotionBench original (test) | Better-GPT52.3 | 2 |