Video Understanding on Aggregate Video Benchmarks Suite
63Overall Average ScoreBaseline
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| BaselineModel=Baseline, Pruning Strength=0%2026.03 | 63 | 67 | 59 | |
| STTSModel=STTS, Pruning Strength=30%2026.03 | 62.8 | 66.7 | 58.9 | |
| Qwen3-VL-4BModel=Qwen3-VL-4B, Pruning Strength=None2026.03 | 62.7 | 63.7 | 61.8 | |
| STTSModel=STTS, Pruning Strength=50%2026.03 | 62.3 | 66.1 | 58.4 | |
| STTSModel=STTS, Pruning Strength=40%2026.03 | 62.1 | 66 | 58.2 | |
| PLM-8BModel=PLM-8B, Pruning Strength=None2026.03 | 61.3 | 68.5 | 54.2 | |
| InternVL3.5-8BModel=InternVL3.5-8B, Pruning Strength=None2026.03 | 60 | 63 | 56.9 | |
| VidGroundFrames=642026.04 | 59.5 | — | — | |
| VidGroundFrames=322026.04 | 58.5 | — | — | |
| TW-GRPOFrames=642026.04 | 58.4 | — | — | |
| Video-RTSFrames=322026.04 | 58 | — | — | |
| Video-RTSFrames=642026.04 | 57.7 | — | — | |
| TW-GRPOFrames=322026.04 | 57.4 | — | — | |
| Qwen2.5-VL-7BFrames=642026.04 | 57.2 | — | — | |
| VidGroundFrames=162026.04 | 56.8 | — | — | |
| TW-GRPOFrames=162026.04 | 56.2 | — | — | |
| Qwen2.5-VL-7BFrames=322026.04 | 56.1 | — | — | |
| Video-RTSFrames=162026.04 | 55.9 | — | — | |
| Qwen2.5-VL-7B-SFTFrames=642026.04 | 55.1 | — | — | |
| Qwen2.5-VL-7BFrames=162026.04 | 54.6 | — | — | |
| LongVILA-R1-7BFrames=322026.04 | 54.1 | — | — | |
| Video-R1Frames=322026.04 | 53.9 | — | — | |
| LongVILA-R1-7BFrames=642026.04 | 53.9 | — | — | |
| Video-R1Frames=642026.04 | 53.3 | — | — | |
| Qwen2.5-VL-7B-SFTFrames=322026.04 | 53.2 | — | — | |
| Video-R1Frames=162026.04 | 52 | — | — | |
| LongVILA-R1-7BFrames=162026.04 | 51.1 | — | — | |
| Qwen2.5-VL-7B-SFTFrames=162026.04 | 50.9 | — | — |