| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Long Video Understanding | LVBench | Accuracy74.8 | 63 | |
| Video Question-Answering | LVBench | Accuracy84.1 | 50 | |
| Video Question Answering | LVBench | Overall Score60.7 | 32 | |
| Video Reasoning | LVBench | LVBench Score43.3 | 24 | |
| Video Understanding | LVBench | Overall Accuracy53.6 | 23 | |
| Video Understanding | LVBench (test) | Accuracy77 | 21 | |
| Video Understanding | LVbench | MAT12.12 | 16 | |
| Long Video Understanding | LVBench (val) | Score58.7 | 15 | |
| Long Video Understanding | LVBench 30-90 min | Accuracy69.2 | 13 | |
| Long-form video understanding and instruction following | LVBench (test) | Accuracy78.4 | 11 | |
| Video Question Answering | LVBench 2024 (test) | Accuracy66.7 | 8 | |
| Extreme Long Video Comprehension | LVBench | Accuracy51.4 | 8 |