Long-video Understanding on Video-MME L
57.7AccuracyVideo-Zero (Peak)
Evaluation Results
| Method | Links | |
|---|---|---|
| Video-Zero (Peak)Backbone=Qwen3-VL-8B-Instruct2026.05 | 57.7 | |
| Video-Zero (Iter 4)Backbone=Qwen3-VL-8B-Instruct2026.05 | 56.6 | |
| V-ZeroBackbone=Qwen3-VL-8B-Instruct2026.05 | 56 | |
| Base ModelBackbone=Qwen3-VL-8B-Instruct2026.05 | 55.7 | |
| VisPlayBackbone=Qwen3-VL-8B-Instruct2026.05 | 54.9 | |
| Video-Zero (Iter 4)Backbone=Qwen3-VL-4B-Instruct2026.05 | 53.6 | |
| Video-Zero (Peak)Backbone=Qwen3-VL-4B-Instruct2026.05 | 53.6 | |
| VisPlayBackbone=Qwen3-VL-4B-Instruct2026.05 | 53.2 | |
| V-ZeroBackbone=Qwen3-VL-4B-Instruct2026.05 | 51.9 | |
| Base ModelBackbone=Qwen3-VL-4B-Instruct2026.05 | 51.6 |