| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Long-video Question Answering | TemporalBench | Binary Accuracy73.2 | 9 | |
| Long video understanding | TemporalBench | Accuracy79.8 | 5 | |
| Video Question Answering | TemporalBench MBA-short QA | Multi-binary Acc36.7 | 5 | |
| Short Video Captioning | TemporalBench (test) | Short Caption Score56.4 | 2 |