| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Long-video understanding | VNBench | Retrieval E Accuracy91.33 | 21 | |
| Video Question Answering | VNBench | Accuracy77.88 | 11 | |
| Long video understanding | VNBench (test) | Retrieval Score0.827 | 7 | |
| Video Understanding | VNBench (val) | Accuracy66.7 | 6 |