Long Video Understanding on Video-MME w/o sub (full)
81.2Score (Long)VideoARM
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| VideoARMBackbone Configuration=o3+gpt-4o2025.12 | 81.2 | 85.8 | 81.3 | 82.8 | |
| VideoARMBackbone Configuration=o3+gpt-4.12025.12 | 75.3 | 86.4 | 78.4 | 80.1 | |
| Gemini-1.5-ProMethod Type=Commercial VLMs2025.12 | 67.4 | 81.7 | 74.3 | 75 | |
| DVDMethod Type=Other Training-Free Approaches2025.12 | 67.3 | — | — | — | |
| GPT-4oMethod Type=Commercial VLMs2025.12 | 65.3 | 80 | 70.3 | 71.9 | |
| AdaReTaKeMethod Type=Open-Source VLMs2025.12 | 65 | 80.6 | 74.9 | 73.5 | |
| OpenAI o3Method Type=Commercial VLMs2025.12 | 63.2 | — | — | — | |
| Gemini-2.0-FlashMethod Type=Commercial VLMs2025.12 | 63 | — | — | — | |
| InternVL2.5-72BMethod Type=Open-Source VLMs2025.12 | 62.6 | 82.8 | 70.9 | 72.1 | |
| MR. VideoMethod Type=Other Training-Free Approaches2025.12 | 61.8 | — | — | — | |
| Logic-in-FramesMethod Type=Other Training-Free Approaches2025.12 | 55.2 | 71.9 | 61.9 | 63 | |
| VideoTreeMethod Type=Other Training-Free Approaches2025.12 | 54.2 | — | — | — | |
| mPLUG-Owl3Method Type=Open-Source VLMs2025.12 | 50.1 | 70 | 57.7 | 59.3 | |
| Qwen2.5-VL-72BMethod Type=Open-Source VLMs2025.12 | — | — | — | 73.3 |