Video Understanding on Video-MME w/o sub (Length Segmented)
80.1Score (Short)VideoLLaMA3-7B
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| VideoLLaMA3-7Bparameter_count=7B2026.06 | 80.1 | 63.7 | 54.9 | 66.2 | |
| VideoChat-Flash-7Bparameter_count=7B2026.06 | 78 | 67.8 | 55.6 | 65.3 | |
| LLaVA-Video-7BBase Model=LLaVA-Video-7B, Input Frames=642026.04 | 76.7 | 62 | 51.4 | 63.4 | |
| Tango†Base Model=LLaVA-Video-7B, Retention Ratio=20%, Intra-LLM=true, Input Frames=642026.04 | 75.3 | 61.1 | 50.4 | 62.3 | |
| TangoBase Model=LLaVA-Video-7B, Retention Ratio=20%, Input Frames=642026.04 | 74.9 | 61.7 | 49.3 | 62 | |
| FastVIDBase Model=LLaVA-Video-7B, Retention Ratio=20%, Input Frames=642026.04 | 74.8 | 60.6 | 50.2 | 61.9 | |
| HoliTomBase Model=LLaVA-Video-7B, Retention Ratio=20%, Input Frames=642026.04 | 74.7 | 60.9 | 50.3 | 62 | |
| Qwen2.5-VL-7BBase Model=Qwen2.5-VL-7B, Input Frames=322026.04 | 74 | 62.7 | 52.4 | 63 | |
| TangoBase Model=LLaVA-Video-7B, Retention Ratio=10%, Input Frames=642026.04 | 73.4 | 59.3 | 49.4 | 60.7 | |
| HoliTomBase Model=LLaVA-Video-7B, Retention Ratio=10%, Input Frames=642026.04 | 73.1 | 58.8 | 48.7 | 60.2 | |
| Tango†Base Model=LLaVA-Video-7B, Retention Ratio=10%, Intra-LLM=true, Input Frames=642026.04 | 73.1 | 59.6 | 49.3 | 60.7 | |
| VisionZipBase Model=LLaVA-Video-7B, Retention Ratio=20%, Input Frames=642026.04 | 73.1 | 59 | 48.9 | 60.3 | |
| InternVL3-8B+MotionEnhancerparameter_count=8B2026.06 | 73 | 62.2 | 51.7 | 62.3 | |
| Qwen2.5-VL-7B+MotionEnhancerparameter_count=7B2026.06 | 72.8 | 61.1 | 50.2 | 61.4 | |
| InternVL3-8B*parameter_count=8B, reproduced=true2026.06 | 71.8 | 61.9 | 52.6 | 62.1 | |
| FastVIDBase Model=LLaVA-Video-7B, Retention Ratio=10%, Input Frames=642026.04 | 71.7 | 58.2 | 48.7 | 59.5 | |
| Qwen2.5-VL-7B*parameter_count=7B, reproduced=true2026.06 | 70.9 | 59.6 | 51.6 | 60.7 | |
| GPT-4V2026.06 | 70.5 | 55.8 | 53.5 | 59.9 | |
| TangoBase Model=Qwen2.5-VL-7B, Retention Ratio=20%, Input Frames=322026.04 | 70.2 | 59.1 | 50.6 | 60 | |
| VisionZipBase Model=Qwen2.5-VL-7B, Retention Ratio=20%, Input Frames=322026.04 | 69.8 | 56.7 | 49.1 | 58.5 | |
| FastVIDBase Model=Qwen2.5-VL-7B, Retention Ratio=20%, Input Frames=322026.04 | 68.9 | 56.8 | 50.3 | 58.7 | |
| VisionZipBase Model=LLaVA-Video-7B, Retention Ratio=10%, Input Frames=642026.04 | 67 | 54.6 | 46.8 | 56.1 | |
| VITA1.5-7Bparameter_count=7B2026.06 | 67 | 54.2 | 47.1 | 56.1 | |
| VidComBase Model=LLaVA-Video-7B, Retention Ratio=10%, Input Frames=642026.04 | 65.6 | 55.7 | 45.9 | 55.7 | |
| FastVBase Model=LLaVA-Video-7B, Retention Ratio=10%, Input Frames=642026.04 | 56.1 | 49.6 | 44.4 | 50 |