| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| LongVideoBench | VideoChat-M1 | Accuracy82.3 | 115 | 1mo ago | |
| LVU | SVT SCALE | Relation Attribute Accuracy76.47 | 44 | 1mo ago | |
| EgoSchema | Accuracy72.2 | 38 | 1mo ago | ||
| LVBench | LLaVA-Video-7B† + Ours (Query-Conditioned Evidential Keyframe Sampling) | Overall Score49.4 | 35 | 9d ago | |
| LVU (test) | S5 + LSMCL | Relation Top-1 Acc67.11 | 16 | 1mo ago | |
| LVU 1.0 (test) | HierarQ | Director Accuracy78.4 | 14 | 1mo ago | |
| Video-MME long-form duration | ColorTrigger | Overall Performance66.1 | 12 | 24d ago | |
| LVBench w/o sub (test) | DVD | Accuracy74.2 | 11 | 26d ago | |
| LV-Bench | MovieChat | ER21.3 | 10 | 1mo ago | |
| LongVideoBench long (val) | VideoSeek | Accuracy73.5 | 7 | 26d ago | |
| VideoMME long subset w/ sub | VideoSeek | Accuracy81.2 | 6 | 26d ago | |
| LongVideoBench | LLaVA-Video | Original Score58.4 | 6 | 1mo ago | |
| LVBench w/ sub (test) | VideoSeek | Accuracy76.7 | 3 | 26d ago |
