Video Understanding Reasoning on VideoMME Medium
73.4AccuracySTD
Evaluation results
| Method | Links | ||
|---|---|---|---|
| STDBackbone=Qwen2-VL-7B-Instruct2025.05 | 73.4 | 1.62 | |
| StreamingBackbone=Qwen2-VL-7B-Instruct2025.05 | 59.2 | 1.36 | |
| STDBackbone=LLaVA-OneVision-7B2025.05 | 52.1 | 1.83 | |
| StreamingBackbone=LLaVA-OneVision-7B2025.05 | 41 | 1.51 | |
| LayerSkipBackbone=LLaVA-OneVision-7B2025.05 | 8.1 | 0.46 | |
| LayerSkipBackbone=Qwen2-VL-7B-Instruct2025.05 | 4.9 | 0.55 |