Real-Time Visual Understanding on StreamingBench (test)
89.5OPHuman
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| HumanSize=-, Frames=-2025.12 | 89.5 | 92 | 93.6 | 91.5 | 95.7 | 92.5 | 88 | 88.8 | 89.7 | 91.3 | 91.5 | |
| Human#Frames=-, RGB (%)=-2026.03 | 89.47 | 92 | 93.6 | 91.47 | 95.65 | 92.52 | 88 | 88.75 | 89.74 | 91.3 | 91.46 | |
| InternVL-3.5-8B#Frames=128, RGB (%)=1002026.03 | 83.47 | 82.03 | 82.65 | 84.62 | 75.47 | 80.06 | 81.48 | 67.89 | 70.45 | 59.04 | 77.2 | |
| InternVL-3.5 + ColorTrigger (Ours, 8B)#Frames=128, RGB (%)=34.3 (↓ 65.7%)2026.03 | 81.57 | 78.91 | 77.29 | 84.94 | 76.1 | 77.57 | 81.48 | 67.48 | 67.61 | 56.91 | 75.24 | |
| Claude 3.5 Sonnet#Frames=20, RGB (%)=1002026.03 | 80.49 | 77.34 | 82.02 | 81.73 | 72.33 | 75.39 | 61.11 | 61.79 | 69.32 | 43.09 | 72.44 | |
| LLaVA-OneVisionSize=7B, Frames=322025.12 | 80.4 | 74.2 | 76 | 80.7 | 72.7 | 71.6 | 67.6 | 65.5 | 65.7 | 45.1 | 71.1 | |
| LLaVA-OneVision-7B#Frames=32, RGB (%)=1002026.03 | 80.38 | 74.22 | 76.03 | 80.72 | 72.67 | 71.65 | 67.59 | 65.45 | 65.72 | 45.08 | 71.12 | |
| TimeChat-Online-7B#Frames=1 fps, RGB (%)=1002026.03 | 80.22 | 82.03 | 79.5 | 83.33 | 76.1 | 78.5 | 78.7 | 64.63 | 69.6 | 57.98 | 75.36 | |
| Gemini 1.5 pro#Frames=1 fps, RGB (%)=1002026.03 | 79.02 | 80.47 | 83.54 | 79.67 | 80 | 84.74 | 77.78 | 64.23 | 71.95 | 48.7 | 75.69 | |
| Qwen2.5-VL-7B#Frames=1 fps, RGB (%)=1002026.03 | 78.32 | 80.47 | 78.86 | 80.45 | 76.73 | 78.5 | 79.63 | 63.41 | 66.19 | 53.19 | 73.68 | |
| LLaVA-NeXTSize=32B, Frames=642025.12 | 78.2 | 70.3 | 73.8 | 76.8 | 63.4 | 69.8 | 57.4 | 56.1 | 64.3 | 38.9 | 67 | |
| GPT-4o#Frames=64, RGB (%)=1002026.03 | 77.11 | 80.47 | 83.91 | 76.47 | 70.19 | 83.8 | 66.67 | 62.19 | 69.12 | 49.22 | 73.28 | |
| Qwen3-VL-8B#Frames=128, RGB (%)=1002026.03 | 76.78 | 75.78 | 79.81 | 80.45 | 71.7 | 77.5 | 77.78 | 68.16 | 65.62 | 56.38 | 73.43 | |
| Qwen3 + ColorTrigger (Ours, 8B)#Frames=128, RGB (%)=34.3 (↓ 65.7%)2026.03 | 75.41 | 74.22 | 78.86 | 79.81 | 75.47 | 77.19 | 77.78 | 62.45 | 64.77 | 54.26 | 72.3 | |
| Qwen2-VLSize=7B, Frames=0.2-1 fps2025.12 | 75.2 | 82.8 | 73.2 | 77.5 | 68.3 | 71 | 72.2 | 61.2 | 61.5 | 46.1 | 69 | |
| InternVL-3.5 + ColorTrigger (Ours, 8B)#Frames=128, RGB (%)=8.1 (↓ 91.9%)2026.03 | 75.07 | 77.34 | 72.56 | 75.64 | 71.7 | 70.4 | 76.85 | 65.04 | 66.48 | 57.98 | 70.72 | |
| Dispider-7B#Frames=1 fps, RGB (%)=1002026.03 | 74.92 | 75.53 | 74.1 | 73.08 | 74.44 | 59.92 | 76.14 | 62.91 | 62.16 | 45.8 | 67.63 | |
| MiniCPM-V2.6Size=8B, Frames=322025.12 | 71.9 | 71.1 | 77.9 | 75.8 | 64.6 | 65.7 | 70.4 | 56.1 | 62.3 | 53.4 | 67.4 | |
| KangarooSize=7B, Frames=642025.12 | 71.1 | 84.4 | 70.7 | 73.2 | 60 | 61.7 | 45.9 | 55.7 | 62 | 38.6 | 62.1 | |
| LongVASize=7B, Frames=1282025.12 | 70 | 63.3 | 61.2 | 70.9 | 62.7 | 59.5 | 61.1 | 53.7 | 54.7 | 34.7 | 60 | |
| InternVL2Size=8B, Frames=162025.12 | 68.1 | 60.9 | 69.4 | 77.1 | 67.7 | 62.9 | 59.3 | 53.3 | 55 | 56.5 | 63.7 | |
| Qwen3 + ColorTrigger (Ours, 8B)#Frames=128, RGB (%)=8.1 (↓ 91.9%)2026.03 | 68.03 | 70.31 | 76.97 | 70.83 | 67.92 | 73.44 | 75.93 | 61.63 | 63.64 | 53.19 | 68.3 | |
| Qwen3-VL-8B grayscale#Frames=128, RGB (%)=02026.03 | 65.57 | 70.31 | 76.34 | 56.41 | 65.41 | 70.31 | 76.85 | 62.45 | 63.64 | 53.19 | 65.61 | |
| InternVL-3.5-8B grayscale#Frames=128, RGB (%)=02026.03 | 60.98 | 78.91 | 60.88 | 55.45 | 65.41 | 63.24 | 75 | 60.16 | 62.22 | 55.85 | 62.08 | |
| Video-LLaMA2Size=7B, Frames=322025.12 | 55.9 | 55.5 | 57.4 | 58.2 | 52.8 | 43.6 | 39.8 | 42.7 | 45.6 | 35.2 | 49.5 | |
| Video-LLaMA2-7B#Frames=32, RGB (%)=1002026.03 | 55.86 | 55.47 | 57.41 | 58.17 | 52.8 | 43.61 | 39.81 | 42.68 | 45.61 | 35.23 | 49.52 | |
| VILA-1.5Size=8B, Frames=142025.12 | 53.7 | 49.2 | 71 | 56.9 | 53.4 | 53.9 | 54.6 | 48.8 | 50.1 | 17.6 | 52.3 | |
| VideoLLM-onlineSize=7B, Frames=0.5 fps2025.12 | 41.9 | 37.6 | 42.6 | 40.5 | 51.5 | 35.5 | 37.7 | 44.9 | 42.7 | 34.6 | 41 | |
| VideoLLM-onlineSize=8B, Frames=2 fps2025.12 | 39.1 | 40.1 | 34.5 | 31.1 | 46 | 32.4 | 31.5 | 34.2 | 42.5 | 27.9 | 36 | |
| VideoLLM-online-8B#Frames=2 fps, RGB (%)=1002026.03 | 39.07 | 40.06 | 34.49 | 31.05 | 45.96 | 32.4 | 31.48 | 34.16 | 42.49 | 27.89 | 35.99 | |
| LLaVA-SFT + OursSize=7B, Frames=0.5 fps2025.12 | 32.6 | 27.1 | 33 | 30.8 | 41.2 | 25 | 27.5 | 34.6 | 33.1 | 24.3 | 31.1 | |
| Flash-VStreamSize=7B, Frames=-2025.12 | 25.9 | 43.6 | 24.9 | 23.9 | 27.3 | 13.1 | 18.5 | 25.2 | 23.9 | 48.7 | 23.2 | |
| Flash-VStream-7B#Frames=-, RGB (%)=1002026.03 | 25.89 | 43.57 | 24.91 | 23.87 | 27.33 | 13.08 | 18.52 | 25.2 | 23.87 | 48.7 | 23.23 |