Video Understanding on VideoMME Overall
73.5AccuracyStreamOV
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| StreamOV#Frames=64, Subtitles=false, Processing Type=Online2026.05 | 73.5 | — | — | |
| Qwen3-Omni-30B#Frames=64, Subtitles=false, Processing Type=Offline2026.05 | 68.5 | — | — | |
| Qwen2.5-Omni-7B#Frames=1fps, Subtitles=false, Processing Type=Offline2026.05 | 64.3 | — | — | |
| Qwen2.5-VL-7B#Frames=1fps, Subtitles=false, Processing Type=Offline2026.05 | 63.2 | — | — | |
| TimeChat-Online-7B#Frames=1fps, Subtitles=false, Processing Type=Online2026.05 | 62.5 | — | — | |
| StreamForest-7B#Frames=1fps, Subtitles=false, Processing Type=Online2026.05 | 61.4 | — | — | |
| Flash-VStream-7B#Frames=1fps, Subtitles=false, Processing Type=Online2026.05 | 61.2 | — | — | |
| LLaVA-OneVision-7B#Frames=32, Subtitles=false, Processing Type=Offline2026.05 | 58.2 | — | — | |
| Dispider-7B#Frames=1fps, Subtitles=false, Processing Type=Online2026.05 | 57.2 | — | — | |
| Kangaroo-7B#Frames=64, Subtitles=false, Processing Type=Offline2026.05 | 56 | — | — | |
| VideoLLM-online-8B#Frames=2fps, Subtitles=false, Processing Type=Online2026.05 | 52.8 | — | — | |
| LongVA-7B#Frames=128, Subtitles=false, Processing Type=Offline2026.05 | 52.6 | — | — | |
| LLaVA-Next-Video-7B#Frames=32, Subtitles=false, Processing Type=Offline2026.05 | 46.6 | — | — | |
| Gemini 1.5 ProAccess=Proprietary2025.12 | — | 75 | 81.3 | |
| GPT-4oAccess=Proprietary2025.12 | — | 65.3 | 77.2 | |
| InternVL2.5Access=Open-Source, Model Scale=7B-8B2025.12 | — | 64.2 | 66.9 | |
| LLaVA-OneVisionAccess=Open-Source, Model Scale=7B-8B2025.12 | — | 58.2 | 61.5 | |
| LongVAAccess=Open-Source, Model Scale=7B-8B2025.12 | — | 52.6 | — | |
| Qwen2.5-VLAccess=Open-Source, Model Scale=7B-8B2025.12 | — | 65.2 | 70.7 | |
| TVG-R1Access=Open-Source, Model Scale=7B-8B2025.12 | — | 64.3 | 69.1 | |
| VideoChat-R1Access=Open-Source, Model Scale=7B-8B2025.12 | — | 64.3 | 69.1 | |
| Zoom-ZeroAccess=Open-Source, Model Scale=7B-8B2025.12 | — | 66 | 71.2 |