Open World Video Understanding on VideoMMMU
71.2Average AccuracyG2F-RAG
Evaluation Results
| Method | Links | |
|---|---|---|
| G2F-RAGSize=14B, Base Model=InternVL3.52026.04 | 71.2 | |
| G2F-RAGSize=8B, Base Model=InternVL3.52026.04 | 68.2 | |
| Qwen3-VLSize=8B2026.04 | 65.3 | |
| InternVL3.5Size=14B2026.04 | 65 | |
| InternVL3.5Size=8B2026.04 | 63.4 | |
| G2F-RAGSize=4B, Base Model=InternVL3.52026.04 | 62.4 | |
| GPT-4o (2024-05-13)Size=-2026.04 | 61.2 | |
| InternVL3.5Size=4B2026.04 | 58.3 | |
| Qwen3-VLSize=4B2026.04 | 56.2 | |
| Gemini1.5 ProSize=-2026.04 | 53.5 | |
| G2F-RAGSize=7B, Base Model=Qwen2.5-VL2026.04 | 53.1 | |
| InternVL3Size=8B2026.04 | 48.9 | |
| Qwen2.5-VLSize=7B2026.04 | 47.4 | |
| VideoLLaMA 3Size=7B2026.04 | 46 | |
| G2F-RAGSize=7B, Base Model=LLaVA-Video2026.04 | 44 | |
| InternVideo2.5Size=7B2026.04 | 43 | |
| Qwen2.5-VLSize=3B2026.04 | 42.3 | |
| LLaVA-VideoSize=7B2026.04 | 36.1 | |
| VILA-1.5Size=8B2026.04 | 20.9 |