Online Video Understanding on OVO-Bench
92.33Backward Tracing Avg.Human Agents
Evaluation results
| Method | Links | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Human Agents2026.02 | 92.33 | 93.96 | 92.57 | 94.83 | 92.7 | 91.09 | 94.02 | 93.2 | 92.59 | 93.02 | 91.37 | 95.48 | 89.67 | 93.56 | 92.9 | 92.81 | |
| Human#Frames=-, Model Type=Baseline2026.03 | 92.3 | 94 | 92.6 | 94.8 | 92.7 | 91.1 | 94 | 93.2 | 92.6 | 93 | 91.4 | 95.5 | 89.7 | 93.6 | 92.9 | 92.8 | |
| Gemini 1.5 ProFrames=1fps2026.02 | 62.54 | 85.91 | 66.97 | 79.31 | 58.43 | 63.37 | 61.96 | 69.32 | 58.59 | 76.35 | 52.64 | 35.53 | 74.24 | 61.67 | 57.15 | 63 | |
| Gemini 1.5 Pro#Frames=1fps, Model Type=Proprietary, Processing Protocol=Online2026.03 | 62.3 | 87.3 | 67 | 80.2 | 54.5 | 68.3 | 67.4 | 70.8 | 68.6 | 75.7 | 52.7 | 35.5 | 74.2 | 61.7 | 57.2 | 65.3 | |
| GPT-4oFrames=642026.02 | 60.75 | 69.8 | 64.22 | 71.55 | 51.12 | 70.3 | 59.78 | 64.46 | 57.91 | 75.68 | 48.66 | 27.58 | 73.21 | 59.4 | 53.4 | 59.54 | |
| GPT-4o#Frames=64, Model Type=Proprietary, Processing Protocol=Online2026.03 | 58.7 | 69.1 | 65.1 | 65.5 | 50 | 68.3 | 63.7 | 63.6 | 49.8 | 71 | 55.4 | 27.6 | 73.2 | 59.4 | 53.4 | 58.6 | |
| EventMemAgentParams=8B, Frames=≤322026.02 | 58.03 | 75.84 | 69.72 | 73.28 | 55.62 | 67.33 | 67.93 | 68.29 | 59.6 | 70.95 | 43.55 | 33.67 | 72.02 | 62.08 | 55.92 | 60.75 | |
| Qwen2-VL-72B# Frames=64, Model Type=Open-source Offline Model2025.12 | 56.95 | 65.77 | 60.55 | 69.83 | 51.69 | 69.31 | 54.35 | 61.92 | 52.53 | 60.81 | 57.53 | 38.83 | 64.07 | 45 | 49.3 | 56.27 | |
| Qwen2-VL-72BParams=72B, Frames=642026.02 | 56.95 | 65.77 | 60.55 | 69.83 | 51.69 | 69.31 | 54.35 | 61.92 | 52.53 | 60.81 | 57.53 | 38.83 | 64.07 | 45 | 49.3 | 56.27 | |
| Em-GardeSize (B)=7B2026.03 | 52.2 | — | — | — | — | — | — | 63 | — | — | — | — | — | — | — | — | |
| StreamForest-7BParams=7B, Frames=1fps2026.02 | 52.02 | 68.46 | 53.21 | 71.55 | 47.75 | 65.35 | 60.87 | 61.2 | 58.92 | 64.86 | 32.26 | 32.81 | 70.59 | 57.08 | 53.49 | 55.57 | |
| StreamForestSize (B)=7B2026.03 | 52 | — | — | — | — | — | — | 61.2 | — | — | — | — | — | — | — | — | |
| Streamo-7B# Frames=2fps*, Model Type=Streamo Framework, Evaluation Frame Rate (FPS)=2, Training Frame Rate (FPS)=12025.12 | 49.18 | 77.18 | 66.06 | 76.72 | 45.51 | 66.34 | 72.83 | 67.44 | 55.56 | 58.11 | 33.87 | 30.84 | 57.55 | 82.5 | 56.96 | 57.86 | |
| Qwen2-VL-7B#Frames=64, Model Type=Open-source, Processing Protocol=Offline2026.03 | 48.6 | 69.1 | 53.2 | 63.8 | 50.6 | 66.3 | 60.9 | 60.7 | 44.4 | 66.9 | 34.4 | 30.1 | 65.7 | 50.8 | 48.9 | 52.7 | |
| Qwen2-VL-7B# Frames=64, Model Type=Open-source Offline Model2025.12 | 46.46 | 60.4 | 50.46 | 56.03 | 47.19 | 66.34 | 55.43 | 55.98 | 47.81 | 35.48 | 56.08 | 31.66 | 65.82 | 48.75 | 48.74 | 50.39 | |
| Qwen2-VL-7BParams=7B, Frames=642026.02 | 46.46 | 60.4 | 50.46 | 56.03 | 47.19 | 66.34 | 55.43 | 55.98 | 47.81 | 35.48 | 56.08 | 31.66 | 65.82 | 48.75 | 48.74 | 50.39 | |
| StreamAgentSize (B)=7B2026.03 | 46.2 | — | — | — | — | — | — | 61.3 | — | — | — | — | — | — | — | — | |
| Qwen3-VL-8BParams=8B, Frames=642026.02 | 46.18 | 75.17 | 58.72 | 72.41 | 57.3 | 70.3 | 59.24 | 65.52 | 56.57 | 69.59 | 12.37 | 47.13 | 67.57 | 52.5 | 55.73 | 55.81 | |
| Streamo-7B# Frames=1fps, Model Type=Streamo Framework2025.12 | 46.1 | 79.19 | 57.8 | 75 | 49.44 | 64.36 | 70.11 | 65.98 | 54.55 | 52.03 | 31.72 | 29.96 | 51.03 | 83.33 | 54.77 | 55.61 | |
| WAT#Frames=1fps, Model Type=Open-source, Processing Protocol=Online2026.03 | 45.2 | 81.2 | 48.3 | 70.7 | 48.3 | 74 | 65.2 | 64.7 | 58.9 | 64.9 | 11.8 | 36.9 | 70.1 | 60.4 | 55.8 | 55.2 | |
| LLaVA-OneVision-7B#Frames=64, Model Type=Open-source, Processing Protocol=Offline2026.03 | 45 | 67.1 | 58.7 | 69.8 | 49.4 | 71.3 | 60.3 | 62.8 | 52.5 | 58.8 | 23.7 | 24.8 | 66.9 | 60.8 | 50.9 | 52.9 | |
| InternVL-V2-8B#Frames=64, Model Type=Open-source, Processing Protocol=Offline2026.03 | 44 | 68.5 | 58.7 | 69 | 44.9 | 67.3 | 56 | 60.7 | 43.1 | 61.5 | 27.4 | 25.8 | 57.6 | 52.9 | 45.4 | 50.1 | |
| LLaVA-OneVision-7B# Frames=64, Model Type=Open-source Offline Model2025.12 | 43.71 | 66.44 | 57.8 | 73.28 | 53.37 | 71.29 | 61.96 | 64.02 | 54.21 | 55.41 | 21.51 | 25.64 | 67.09 | 58.75 | 50.5 | 52.74 | |
| LLaVA-OneVision-7BParams=7B, Frames=642026.02 | 43.71 | 66.44 | 57.8 | 73.28 | 53.37 | 71.29 | 61.96 | 64.02 | 54.21 | 55.41 | 21.51 | 25.64 | 67.09 | 58.75 | 50.5 | 52.74 | |
| InternVL-V2-8B# Frames=64, Model Type=Open-source Offline Model2025.12 | 43.44 | 67.11 | 60.55 | 63.79 | 46.07 | 68.32 | 56.52 | 60.39 | 48.15 | 57.43 | 24.73 | 26.5 | 59.14 | 54.14 | 46.6 | 50.15 | |
| InternVL-V2-8BParams=8B, Frames=642026.02 | 43.44 | 67.11 | 60.55 | 63.79 | 46.07 | 68.32 | 56.52 | 60.39 | 48.15 | 57.43 | 24.73 | 26.5 | 59.14 | 54.14 | 46.6 | 50.15 | |
| Streamo-3B# Frames=1fps, Model Type=Streamo Framework2025.12 | 41.76 | 78.52 | 52.29 | 67.24 | 44.38 | 55.45 | 71.2 | 61.51 | 51.18 | 57.43 | 16.67 | 27.94 | 50.72 | 82.5 | 53.72 | 52.33 | |
| StreamAgentParams=7B, Frames=1fps2026.02 | 41.7 | 71.2 | 53.2 | 63.6 | 53.9 | 67.3 | 58.7 | 61.3 | 54.8 | 58.1 | 25.8 | 35.9 | 48.4 | 52 | 45.4 | 49.4 | |
| LLaVA-NeXT-Video-7B#Frames=64, Model Type=Open-source, Processing Protocol=Offline2026.03 | 41.7 | 69.8 | 59.6 | 66.4 | 50.6 | 72.3 | 61.4 | 63.3 | 51.2 | 64.2 | 9.7 | 34.1 | 67.6 | 60.8 | 54.2 | 53.1 | |
| TimeChat-Online-7B#Frames=1fps, Model Type=Open-source, Processing Protocol=Online2026.03 | 41.7 | 75.2 | 46.8 | 70.7 | 47.8 | 69.3 | 61.4 | 61.9 | 55.9 | 59.5 | 9.7 | 31.6 | 38.5 | 40 | 36.7 | 46.7 | |
| LLaVA-Video-7B# Frames=64, Model Type=Open-source Offline Model2025.12 | 40.4 | 69.13 | 58.72 | 68.83 | 49.44 | 74.26 | 59.78 | 63.52 | 56.23 | 57.43 | 7.53 | 34.1 | 69.95 | 60.42 | 54.82 | 52.91 | |
| LLaVA-Video-7BParams=7B, Frames=642026.02 | 40.4 | 69.13 | 58.72 | 68.83 | 49.44 | 74.26 | 59.78 | 63.52 | 56.23 | 57.43 | 7.53 | 34.1 | 69.95 | 60.42 | 54.82 | 52.91 | |
| LongVU-7B#Frames=1fps, Model Type=Open-source, Processing Protocol=Offline2026.03 | 39.5 | 55.7 | 49.5 | 59.5 | 48.3 | 68.3 | 63 | 57.4 | 43.1 | 66.2 | 9.1 | 16.6 | 69 | 60 | 48.5 | 48.5 | |
| ET-Instruct-3B†# Frames=1fps, Model Type=Streamo Framework Adaptation, Offline Support Data (LLaVA-Video)=true2025.12 | 36.22 | 71.14 | 50.46 | 67.24 | 37.08 | 60.4 | 60.33 | 57.78 | 48.82 | 48.56 | 11.29 | 13.68 | 48.62 | 60 | 40.77 | 44.92 | |
| DispiderSize (B)=7B2026.03 | 36.1 | — | — | — | — | — | — | 54.6 | — | — | — | — | — | — | — | — | |
| Dispider-7B# Frames=1fps, Model Type=Open-source Online Model2025.12 | 36.06 | 57.72 | 49.54 | 62.07 | 44.94 | 61.39 | 51.63 | 54.55 | 48.48 | 55.41 | 4.3 | 18.05 | 37.36 | 48.75 | 34.72 | 41.78 | |
| DispiderParams=7B, Frames=1fps2026.02 | 36.06 | 57.72 | 49.54 | 62.07 | 44.94 | 61.39 | 51.63 | 54.55 | 48.48 | 55.41 | 4.3 | 18.05 | 37.36 | 48.75 | 34.72 | 41.78 | |
| LongVU-7B# Frames=1fps, Model Type=Open-source Offline Model2025.12 | 35.01 | 53.69 | 53.21 | 62.93 | 47.75 | 68.32 | 59.78 | 57.61 | 40.74 | 59.46 | 4.84 | 12.18 | 69.48 | 60.83 | 47.5 | 46.71 | |
| LongVU-7BParams=7B, Frames=1fps2026.02 | 35.01 | 53.69 | 53.21 | 62.93 | 47.75 | 68.32 | 59.78 | 57.61 | 40.74 | 59.46 | 4.84 | 12.18 | 69.48 | 60.83 | 47.5 | 46.71 | |
| ET-Instruct-3B# Frames=1fps, Model Type=Streamo Framework Adaptation2025.12 | 28.52 | 65.1 | 35.78 | 56.9 | 35.39 | 24.75 | 60.87 | 46.47 | 41.81 | 35.14 | 8.6 | 20.06 | 52.31 | 67.5 | 46.62 | 40.54 | |
| Flash-VStream-7B# Frames=1fps, Model Type=Open-source Online Model2025.12 | 27.38 | 24.16 | 29.36 | 28.45 | 33.71 | 25.74 | 28.8 | 28.37 | 39.06 | 37.16 | 5.91 | 8.02 | 67.25 | 60 | 45.09 | 33.61 | |
| Flash-VStream-7BParams=7B, Frames=1fps2026.02 | 27.38 | 24.16 | 29.36 | 28.45 | 33.71 | 25.74 | 28.8 | 28.37 | 39.06 | 37.16 | 5.91 | 8.02 | 67.25 | 60 | 45.09 | 33.61 | |
| Flash-VStream-7B#Frames=1fps, Model Type=Open-source, Processing Protocol=Online2026.03 | 25.4 | 25.5 | 32.1 | 29.3 | 33.7 | 29.7 | 28.8 | 29.9 | 36.4 | 33.8 | 5.9 | 5.4 | 67.3 | 60 | 44.2 | 33.2 | |
| FVStreamSize (B)=7B2026.03 | 25.4 | — | — | — | — | — | — | 29.9 | — | — | — | — | — | — | — | — | |
| VideoLLM-online-8B# Frames=2fps, Model Type=Open-source Online Model2025.12 | 17.73 | 8.05 | 23.85 | 12.07 | 14.04 | 45.54 | 21.2 | 20.79 | 22.22 | 18.8 | 12.18 | — | — | — | — | — | |
| VideoLLM-online-8BParams=8B, Frames=2fps2026.02 | 17.73 | 8.05 | 23.85 | 12.07 | 14.04 | 45.54 | 21.2 | 20.79 | 22.22 | 18.8 | 12.18 | — | — | — | — | — | |
| VideoLLM-online-8B#Frames=2fps, Model Type=Open-source, Processing Protocol=Online2026.03 | 17.7 | 8.1 | 23.9 | 12.1 | 14 | 45.5 | 21.2 | 20.8 | 22.2 | 18.8 | 12.2 | — | — | — | — | — | |
| VideoLLM-OnlineSize (B)=8B2026.03 | 17.7 | — | — | — | — | — | — | 20.8 | — | — | — | — | — | — | — | — | |
| DispiderSize=7B, Frames=1 fps2026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 41.8 | |
| Flash-VStreamSize=7B, Frames=1 fps2026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 33.6 | |
| FluxMemSize=7B, Frames=1 fps2026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 53.3 | |
| LLaVA-OneVisionSize=7B, Frames=322026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 52.9 | |
| LLaVA-VideoSize=7B, Frames=1 fps2026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 53.1 | |
| LongVUSize=7B, Frames=1 fps2026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 48.5 | |
| Qwen2.5-VL†Size=7B, Frames=1 fps2026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 49.8 | |
| StreamForestSize=7B, Frames=1 fps2026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 55.6 | |
| TimeChat-OnlineSize=7B, Frames=1 fps2026.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 47.6 |