Long-video Question Answering on MMBench-Video (val)
1.97ScoreInternVL2.5-78B
Evaluation Results
| Method | Links | |
|---|---|---|
| InternVL2.5-78B2025.01 | 1.97 | |
| GPT-4o2025.01 | 1.87 | |
| Tarsier2-7BNumber of frames=128f2025.01 | 1.82 | |
| LLaVA-Video-72B2025.01 | 1.71 | |
| Qwen2-VL-72B2025.01 | 1.7 | |
| InternVL2.5-8B2025.01 | 1.68 | |
| VILA-1.5-40B2025.01 | 1.61 | |
| LLaVA-Video-7B2025.01 | 1.6 | |
| Tarsier-34B2025.01 | 1.46 | |
| Qwen2-VL-7B2025.01 | 1.44 | |
| Gemini-1.5-Pro2025.01 | 1.3 |