Long Video Understanding on Video-MME Overall
87AccuracyGemini-2.5-Pro
Evaluation results
| Method | Links | |
|---|---|---|
| Gemini-2.5-ProInput Frames=>=642025.12 | 87 | |
| AVPBackbone=Gemini-2.5-Pro2025.12 | 85.3 | |
| Gemini-2.5-Pro2025.12 | 82.4 | |
| LVAgentAuxiliary Subtitles=true2025.12 | 81.7 | |
| AVPBackbone=Gemini-2.5-Flash2025.12 | 81.2 | |
| Qwen-3-VL2025.12 | 79.2 | |
| Seed-1.5-VL2025.12 | 77.9 | |
| Seed1.5-VLSize=8B, Input Frames=>=642025.12 | 77.9 | |
| GPT-4oFrames=1fps2026.06 | 77.2 | |
| Gemini-2.5-Flash2025.12 | 74.2 | |
| SiLVR2025.12 | 74.1 | |
| VideoMem*Size=8B, Input Frames=642025.12 | 73.6 | |
| AdaReTaKe2025.12 | 73.5 | |
| Qwen2.5-VLSize=72B, Input Frames=>=642025.12 | 73.3 | |
| VideoLucy2025.12 | 72.5 | |
| Eagle2.5Size=8B, Input Frames=>=642025.12 | 72.4 | |
| GPT-4.12025.12 | 72 | |
| GPT-4o2025.12 | 71.9 | |
| GPT-4oInput Frames=>=642025.12 | 71.9 | |
| Vgent2025.12 | 68.9 | |
| Qwen3-VL*Size=8B, Input Frames=642025.12 | 67.9 | |
| Thinking-QwenVLFrames=1fps (100%)2026.04 | 67.7 | |
| VILAMPSize=7B, Frames=1fps2026.06 | 67.5 | |
| Video-XL2Size=8B, Input Frames=>=642025.12 | 66.6 | |
| LongVITA-128KSize=14B, Frames=2562026.06 | 66.4 | |
| VideoLLaMA3Size=7B, Input Frames=>=642025.12 | 66.2 | |
| InternVL3.5Size=8B, Input Frames=>=642025.12 | 66 | |
| VideoChat-FlashSize=7B, Frames=5122026.06 | 65.3 | |
| Qwen2.5-VLSize=7B, Input Frames=>=642025.12 | 65.1 | |
| LongVILA-R1Size=7B, Frames=512, Context=132k2026.06 | 65.1 | |
| Flow4AgentSize=7B, Input Frames=>=642025.12 | 64.7 | |
| NVILASize=8B, Input Frames=>=642025.12 | 64.2 | |
| VITALSize=7B, Frames=1024, Context=32k2026.06 | 64.1 | |
| CoVERSize=7B, Frames=768L + 32H, Context=16k2026.06 | 63.8 | |
| Hour-LLaVASize=7B, Frames=1fps2026.06 | 63.6 | |
| Qwen2.5-VL-7BFrames=1fps2026.04 | 63.2 | |
| Qwen2.5-VL*Size=7B, Frames=128, Context=16k2026.06 | 63.1 | |
| Video-RTS2025.12 | 63 | |
| LLaVA-videoSize=7B, Input Frames=>=642025.12 | 62.6 | |
| TimeChat-Online-7BFrames=1fps (100%)2026.04 | 62.4 | |
| Video-RAGSize=7B2026.06 | 62.1 | |
| ApolloSize=7B, Input Frames=>=642025.12 | 61.3 | |
| FrameMind2025.12 | 60.9 | |
| GPT-4VInput Frames=>=642025.12 | 60.7 | |
| GPT-4VFrames=1fps2026.06 | 60.7 | |
| LongVU2025.12 | 60.6 | |
| VideoTree2025.12 | 60.6 | |
| LongVUSize=7B, Frames=1fps2026.06 | 60.6 | |
| ConanSize=7B, Frames=322026.06 | 60.5 | |
| LongVILASize=7B, Frames=2562026.06 | 60.1 | |
| Video-XL-ProSize=7B, Frames=2402026.06 | 60 | |
| Video-R1Size=7B, Frames=32, Context=4k2026.06 | 59.3 | |
| Video-MTRSize=7B, Frames=32, Context=4k2026.06 | 59 | |
| LLaVA-OnevisionSize=7B, Input Frames=>=642025.12 | 58.2 | |
| LLaVA-OneVisionSize=7B, Frames=322026.06 | 58.2 | |
| VideoMindSize=7B2026.06 | 58.2 | |
| VambaSize=10B, Frames=10242026.06 | 57.8 | |
| Dispider-7BFrames=1fps2026.04 | 57.2 | |
| Thinking-QwenVLFrames=1fps (↓93.75%)2026.04 | 56.3 | |
| Kangaroo-7BFrames=642026.04 | 56 | |
| Video-XL-7BFrames=1282026.04 | 55.5 | |
| VISTA-7BFrames=-2026.04 | 55.5 | |
| Video-XLSize=7B, Frames=2562026.06 | 55.5 | |
| VideoChat2Size=7B, Frames=162026.06 | 54.6 | |
| LongVASize=7B, Input Frames=>=642025.12 | 54.3 | |
| LongVASize=7B, Frames=1282026.06 | 54.3 | |
| Video-CCAM-14BFrames=962026.04 | 53.2 | |
| VideoChat-Online-8BFrames=2fps2026.04 | 52.8 | |
| LongVA-7BFrames=1282026.04 | 52.6 | |
| VideoLLaMA2Size=7B, Input Frames=>=642025.12 | 46.6 | |
| LLaVA-NeXT-Video-7BFrames=322026.04 | 46.6 | |
| VideoLLaMA2Size=7B, Frames=322026.06 | 46.6 | |
| LLaVA-Next-VideoSize=7B, Input Frames=>=642025.12 | 46.5 | |
| LLaVA-NeXT-VideoSize=7B, Frames=322026.06 | 46.5 | |
| VideoChat-TSize=7B, Frames=122026.06 | 46.3 | |
| ShareGPT4VideoSize=8B, Input Frames=>=642025.12 | 43.6 | |
| ShareGPT4VideoSize=8B, Frames=162026.06 | 43.6 | |
| Video-LLaVASize=7B, Input Frames=>=642025.12 | 40.4 | |
| Video-LLaVASize=7B, Frames=82026.06 | 40.4 | |
| VideoChat2-7BFrames=162026.04 | 39.5 | |
| MovieChat-7BFrames=20482026.04 | 38.2 |