Video Understanding on MLVU (MAT and Wall-time Speedup)
12.42MATHIPPO
Evaluation results
| Method | Links | ||
|---|---|---|---|
| HIPPOBackbone Model=Video-SALMONN2+ 7&72B2026.01 | 12.42 | 2.78 | |
| HIPPOBackbone Model=LLaVA-OneVision 7&72B2026.01 | 7.96 | 3.17 | |
| HIPPOBackbone Model=Qwen2.5-VL 7&72B2026.01 | 7.19 | 2.54 | |
| SpecVLMBackbone Model=Video-SALMONN2+ 7&72B2026.01 | 4 | 2.13 | |
| SD-treeBackbone Model=Video-SALMONN2+ 7&72B2026.01 | 3.86 | 1.83 | |
| SD-treeBackbone Model=Qwen2.5-VL 7&72B2026.01 | 3.84 | 2.01 | |
| Vanilla SDBackbone Model=Qwen3-VL 7&32B2026.01 | 3.67 | 1.23 | |
| SpecVLMBackbone Model=Qwen2.5-VL 7&72B2026.01 | 3.48 | 2.14 | |
| Vanilla SDBackbone Model=Qwen2.5-VL 7&72B2026.01 | 3.39 | 1.64 | |
| Vanilla SDBackbone Model=Video-SALMONN2+ 7&72B2026.01 | 3.3 | 1.53 | |
| SD-treeBackbone Model=LLaVA-OneVision 7&72B2026.01 | 3.3 | 1.9 | |
| SpecVLMBackbone Model=LLaVA-OneVision 7&72B2026.01 | 2.96 | 2 | |
| Vanilla SDBackbone Model=LLaVA-OneVision 7&72B2026.01 | 2.87 | 1.64 | |
| HIPPOBackbone Model=Qwen3-VL 7&32B2026.01 | 2.82 | 2.15 | |
| SD-treeBackbone Model=Qwen3-VL 7&32B2026.01 | 2.51 | 1.53 | |
| SpecVLMBackbone Model=Qwen3-VL 7&32B2026.01 | 2.33 | 1.85 |