Streaming Video Understanding on OVOBench Realtime
93.2Average ScoreHuman
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| HumanSize=-, Frames=-2026.03 | 93.2 | 94 | 92.6 | 94.8 | 92.7 | 91.1 | 94 | |
| Streamo-4BCategory=Streamo Framework, Model Size=4B, Base Model=Qwen3-VL-4B2025.12 | 72.2 | — | — | — | — | — | — | |
| Gemini-1.5-proCategory=Proprietary Models2025.12 | 69.3 | — | — | — | — | — | — | |
| Gemini 1.5 ProSize=-, Frames=1 fps, Type=Proprietary2026.03 | 69.3 | 85.9 | 67 | 79.3 | 58.4 | 63.4 | 62 | |
| FluxMemSize=7B, Frames=1 fps, Strategy=Online, Training=Training-Free2026.03 | 67.2 | 81.2 | 59.6 | 70.7 | 53.4 | 75.2 | 63 | |
| Qwen3-VL-4BCategory=Streamo Framework, Model Size=4B2025.12 | 66.5 | — | — | — | — | — | — | |
| ViSpeakSize=7B, Frames=1 fps, Strategy=Online, Training=Training-Based2026.03 | 66.3 | 75.2 | 58.7 | 71.6 | 51.1 | 74.3 | 66.9 | |
| Streamo-7BCategory=Streamo Framework, Model Size=7B, Base Model=Qwen2.5-VL-7B2025.12 | 66 | — | — | — | — | — | — | |
| CausalMemFrames=0.5fps2026.06 | 65.7 | 71.8 | 61.5 | 76.7 | 48.9 | 76.2 | 66.8 | |
| GPT-4oCategory=Proprietary Models2025.12 | 64.5 | — | — | — | — | — | — | |
| GPT-4oSize=-, Frames=64, Type=Proprietary2026.03 | 64.5 | 69.8 | 64.2 | 71.6 | 51.1 | 70.3 | 59.8 | |
| LLaVA-OneVisionSize=7B, Frames=64/32, Strategy=Offline2026.03 | 64 | 66.4 | 57.8 | 73.3 | 53.4 | 71.3 | 62 | |
| VisionZipFrames=0.5fps, method_property=off-line2026.06 | 63.7 | 68.5 | 55 | 77.6 | 48.3 | 74.3 | 65.2 | |
| StreamingTomFrames=0.5fps2026.06 | 63.6 | 67.8 | 61.5 | 75.9 | 47.8 | 68.3 | 66.3 | |
| LLaVA-VideoSize=7B, Frames=64, Strategy=Offline2026.03 | 63.5 | 69.1 | 58.7 | 68.8 | 49.4 | 74.3 | 59.8 | |
| Qwen2.5-VL†Size=7B, Frames=1 fps, Strategy=Online, Training=Training-Free2026.03 | 63.3 | 79.2 | 53.2 | 67.2 | 51.7 | 71.3 | 57.1 | |
| LLaVA-OneVisionFrames=322026.06 | 62.6 | 67.1 | 61.5 | 74.1 | 45.5 | 69.3 | 65.2 | |
| StreamingVLM-7BCategory=Open-source Online Models, Model Size=7B2025.12 | 62 | — | — | — | — | — | — | |
| LiveVLMFrames=0.5fps2026.06 | 61.6 | 67.1 | 60.5 | 70.7 | 46.1 | 71.3 | 61.9 | |
| Streamo-3BCategory=Streamo Framework, Model Size=3B, Base Model=Qwen2.5-VL-3B2025.12 | 61.5 | — | — | — | — | — | — | |
| TimeChat-OnlineSize=7B, Frames=1 fps, Strategy=Online, Training=Training-Based2026.03 | 61.4 | 74.5 | 48.6 | 68.1 | 48.3 | 69.3 | 59.8 | |
| StreamForestSize=7B, Frames=1 fps, Strategy=Online, Training=Training-Based2026.03 | 61.2 | 68.5 | 53.2 | 71.6 | 47.8 | 65.4 | 60.9 | |
| Streamo-2BCategory=Streamo Framework, Model Size=2B, Base Model=InternVL3-2B2025.12 | 61.1 | — | — | — | — | — | — | |
| InternVL2Size=8B, Frames=64/16, Strategy=Offline2026.03 | 60.4 | 67.1 | 60.6 | 63.8 | 46.1 | 68.3 | 56.5 | |
| InternVL3-2BCategory=Streamo Framework, Model Size=2B2025.12 | 59.5 | — | — | — | — | — | — | |
| Qwen2.5-VL-7BCategory=Streamo Framework, Model Size=7B2025.12 | 58.8 | — | — | — | — | — | — | |
| LongVUSize=7B, Frames=1 fps, Strategy=Offline2026.03 | 57.4 | 55.7 | 49.5 | 59.5 | 48.3 | 68.3 | 63 | |
| ReKVFrames=0.5fps2026.06 | 57.3 | — | — | — | — | — | — | |
| Dispider-7BCategory=Open-source Online Models, Model Size=7B2025.12 | 54.6 | — | — | — | — | — | — | |
| Qwen2.5-VL-3BCategory=Streamo Framework, Model Size=3B2025.12 | 54.6 | — | — | — | — | — | — | |
| DispiderSize=7B, Frames=1 fps, Strategy=Online, Training=Training-Based2026.03 | 54.6 | 57.7 | 49.5 | 62.1 | 44.9 | 61.4 | 51.6 | |
| Flash-VStream-7BCategory=Open-source Online Models, Model Size=7B2025.12 | 28.4 | — | — | — | — | — | — | |
| Flash-VStreamSize=7B, Frames=1 fps, Strategy=Online, Training=Training-Based2026.03 | 28.4 | 24.2 | 29.4 | 28.5 | 33.7 | 25.7 | 28.8 | |
| Flash-VStream-7BFrames=1 fps2026.01 | 28.37 | 24.16 | 29.36 | 28.45 | 33.71 | 25.74 | 28.8 | |
| Event-VStreamFrames=2 fps2026.01 | 28.15 | 24.38 | 29.36 | 27.59 | 25.09 | 33.66 | 28.8 | |
| VideoLLM-online-8BCategory=Open-source Online Models, Model Size=8B2025.12 | 20.8 | — | — | — | — | — | — | |
| VideoLLM-OnlineSize=8B, Frames=2 fps, Strategy=Online, Training=Training-Based2026.03 | 20.8 | 8.1 | 23.9 | 12.1 | 14 | 45.5 | 21.2 | |
| VideoLLM-online-8BFrames=2 fps2026.01 | 17.73 | 8.05 | 23.85 | 12.07 | 14.04 | 45.54 | 18.8 |