| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| LongVideoBench (test) | Accuracy66.7 | 28 | 4mo ago | ||
| VideoMME | Accuracy (w subs)81.3 | 13 | 4mo ago | ||
| MLVU MCQ (val) | Baseline | Score70.3 | 7 | 4mo ago | |
| LongVideo (val) | Qwen3-VL-4B | Score62.8 | 7 | 4mo ago | |
| VideoMME Sub (test) | Qwen3-VL-4B | Score74 | 7 | 4mo ago | |
| VideoMME (test) | Qwen3-VL-4B | Score69.3 | 7 | 4mo ago | |
| VideoMME 32K context slice | Qwen3.5-9B | Score70 | 4 | 1mo ago | |
| VideoMME 16K context slice | Qwen3.5-9B | Score66 | 4 | 1mo ago | |
| LongVideo Sub (val) | Baseline | Score60.9 | 4 | 4mo ago |