| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Long Video Understanding | LongVideoBench (val) | Accuracy80 | 139 | |
| Long Video Understanding | LongVideoBench | Score66.7 | 110 | |
| Long-form Video Understanding | LongVideoBench | Accuracy74.4 | 82 | |
| Video Understanding | LongVideoBench | LongVideoBench Score59.2 | 79 | |
| Video Understanding | LongVideoBench 1-60min | Accuracy56.8 | 49 | |
| Video Question Answering | LongVideoBench (LVB) 58 (test) | Accuracy66.42 | 45 | |
| Video Question-Answering | LongVideoBench | Accuracy68 | 34 | |
| Long-context video understanding | LongVideoBench (test) | Accuracy66.7 | 28 | |
| Video Reasoning | LongVideoBench | LongVideoBench Score59 | 24 | |
| Video Understanding | LongVideoBench (test) | Accuracy (8-15s)77.78 | 21 | |
| Video Question Answering | LongVideoBench KFS-Bench (full set) | Accuracy65.5 | 20 | |
| Long Video Understanding | LongVideoBench 23sec-60 min | Accuracy74.4 | 19 | |
| Long Video Understanding | LongVideoBench (LongVB) | Accuracy69.2 | 17 | |
| Video Question Answering | LongVideoBench (val) | Overall Score66.7 | 12 | |
| Video Question Answering | LongVideoBench (test) | Accuracy59.6 | 9 | |
| Speculative Decoding | LongVideoBench ~15k visual tokens | Tau (τ)3.82 | 8 | |
| Long Video Understanding | LongVideoBench Long | Accuracy70 | 7 | |
| Video Question Answering | LongVideoBench 1.0 (test) | Accuracy67.2 | 6 | |
| Video Understanding | LongVideoBench (LVB) | Overall Score65 | 5 | |
| Long Video Understanding | LongVideoBench (test) | Accuracy66.7 | 4 | |
| General Video Understanding | LongVideoBench | Accuracy52.13 | 2 | |
| Long Video Understanding | LongVideoBench | Pearson's r-0.129 | 1 |