Dialogue Evaluation on SVBench
58.26SAGPT-4o
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| GPT-4otype=Closed-source LVLM2026.06 | 58.26 | 64.76 | 70.75 | 67.68 | 55.82 | 62.57 | 61.27 | |
| GPT-4Vtype=Closed-source LVLM2026.06 | 56.03 | 62.61 | 69.09 | 65.36 | 53.73 | 60.3 | 59.12 | |
| LiveStarfine-tuning on the SVBench training set=true2026.06 | 54.06 | 61.08 | 66.43 | 66.06 | 52.67 | 58.95 | 57.41 | |
| Qwen2.5-VLtype=Open-source Video-LLM2026.06 | 52.54 | 59.85 | 65.52 | 64.64 | 51.23 | 57.57 | 55.21 | |
| MiniCPM-V 2.6type=Open-source Video-LLM2026.06 | 51.7 | 59.5 | 65.33 | 61.72 | 50.09 | 56.63 | 53.15 | |
| LiveStarProtype=Online Assistant2026.06 | 47.49 | 56.58 | 63.86 | 59.7 | 48.92 | 53.72 | 52.2 | |
| InternVideo2.5type=Open-source Video-LLM2026.06 | 46.83 | 53.48 | 58.22 | 58.91 | 47.02 | 51.73 | 49.16 | |
| LiveStartype=Online Assistant2026.06 | 46.43 | 53.75 | 59.36 | 57.29 | 45.64 | 51.37 | 49.76 | |
| InternVL2.5type=Open-source Video-LLM2026.06 | 43.73 | 50.7 | 56.61 | 55.03 | 43.46 | 48.73 | 46.89 | |
| LLaVA-NeXT-Videotype=Open-source Video-LLM2026.06 | 37.71 | 44.59 | 52.05 | 41.8 | 36.58 | 41.4 | 38.76 |