Visual Spatial Intelligence on VSI-Bench
79.2Average ScoreHuman Level†
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Human Level†2026.03 | 79.2 | 94.3 | 47 | 60.4 | 45.9 | 94.7 | 95.8 | 95.8 | 100 | — | |
| OmniStream-7B2026.03 | 70.6 | 73.2 | 55.7 | 76.9 | 74.8 | 72.3 | 82.1 | 45.4 | 84.6 | — | |
| SpaceMind2026.03 | 69.6 | 73.3 | 61.4 | 77.3 | 74.2 | 67.2 | 88.4 | 44.3 | 70.5 | — | |
| InternVideo3Model Category=Open-Weight Models2026.06 | 68.1 | — | — | — | — | — | — | — | — | — | |
| Full PrecisionModel=InternVL3.5-38B-Instruct, Weight bits=16 bit, Memory=71.51 GB2026.07 | 66.3 | — | — | — | — | — | — | — | — | — | |
| Cog3DMap-8BModel Category=Spatial-Enhanced Models2026.03 | 65.1 | 69.6 | 54.8 | 67.8 | 67.1 | 64.8 | 85.6 | 43 | 67.9 | — | |
| Pixelis (Qwen3-VL-8B-Instruct)Size=8B2026.03 | 64.4 | — | — | — | — | — | — | — | — | — | |
| InternVL3.5-A3B (with tools)Size=30B2026.03 | 63.7 | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL-30B-A3B-InstructSize=30B2026.03 | 63.2 | — | — | — | — | — | — | — | — | — | |
| Pixelis: SFT + RFTSize=8B2026.03 | 62.8 | — | — | — | — | — | — | — | — | — | |
| Pixelis: SFT + TTRLSize=8B2026.03 | 62.8 | — | — | — | — | — | — | — | — | — | |
| PRM (process reward; tools; 8B)Size=8B2026.03 | 62.3 | — | — | — | — | — | — | — | — | — | |
| Pixelis: RFT + TTRLSize=8B2026.03 | 61.9 | — | — | — | — | — | — | — | — | — | |
| VST-7BModel Category=Spatial-Enhanced Models2026.03 | 61.2 | 71.6 | 43.8 | 75.5 | 69.2 | 60 | 55.6 | 44.3 | 69.2 | — | |
| VLM-3R-7B2026.03 | 60.9 | 70.2 | 49.4 | 69.2 | 67.1 | 65.4 | 80.5 | 45.4 | 40.1 | — | |
| VLM-3R-7BModel Category=Spatial-Enhanced Models2026.03 | 60.9 | 70.2 | 49.4 | 69.2 | 67.1 | 65.4 | 80.5 | 45.4 | 40.1 | — | |
| Pixelis: RFT onlySize=8B2026.03 | 60.8 | — | — | — | — | — | — | — | — | — | |
| Full PrecisionModel=InternVL3.5-14B-Instruct, Weight bits=16 bit, Memory=28.16 GB2026.07 | 60.8 | — | — | — | — | — | — | — | — | — | |
| VST-7B2026.03 | 60.6 | 72 | 44.4 | 74.3 | 68.3 | 59.7 | 55.8 | 44.9 | 65.2 | — | |
| Pixel Reasoner (Qwen3-VL-8B-Instruct)Size=8B2026.03 | 60.1 | — | — | — | — | — | — | — | — | — | |
| Pixelis: SFT onlySize=8B2026.03 | 60.1 | — | — | — | — | — | — | — | — | — | |
| Pixelis: TTRL onlySize=8B2026.03 | 59.6 | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL-8B-InstructSize=8B2026.03 | 59.4 | — | — | — | — | — | — | — | — | — | |
| Qwen3-VL-8BModel Category=Open-Weight Models, Parameter count=8B2026.06 | 59.1 | — | — | — | — | — | — | — | — | — | |
| Full PrecisionModel=InternVL3.5-8B-Instruct, Weight bits=16 bit, Memory=15.89 GB2026.07 | 56.3 | — | — | — | — | — | — | — | — | — | |
| GPTQModel=InternVL3.5-38B-Instruct, Weight bits=3 bit, Memory=12.71 GB2026.07 | 56.12 | — | — | — | — | — | — | — | — | — | |
| InternVL3.5-8BModel Category=Open-Weight Models, Parameter count=8B2026.06 | 56 | — | — | — | — | — | — | — | — | — | |
| GPTQModel=InternVL3.5-14B-Instruct, Weight bits=3 bit, Memory=5.38 GB2026.07 | 52.6 | — | — | — | — | — | — | — | — | — | |
| VG-LLM-8BModel Category=Spatial-Enhanced Models2026.03 | 50.7 | 67.9 | 37.7 | 58.6 | 62 | 46.6 | 40.7 | 32.4 | 59.2 | — | |
| GPTQModel=InternVL3.5-8B-Instruct, Weight bits=3 bit, Memory=2.83 GB2026.07 | 49.01 | — | — | — | — | — | — | — | — | — | |
| Spatial-MLLM-4BModel Category=Spatial-Enhanced Models2026.03 | 48.4 | 65.3 | 34.8 | 63.1 | 45.1 | 41.3 | 46.2 | 33.5 | 46.3 | — | |
| Spatial-MLLM-4BModel Category=Open-source, Number of Frames=162025.05 | 48.4 | 65.3 | 34.8 | 63.1 | 45.1 | 41.3 | 46.2 | 33.5 | 46.3 | — | |
| VG-LLM-4BModel Category=Spatial-Enhanced Models2026.03 | 47.3 | 66 | 37.8 | 55.2 | 59.2 | 44.6 | 45.6 | 33.5 | 36.4 | — | |
| 3DRS-7BModel Category=Spatial-Enhanced Models2026.03 | 45.9 | 68.7 | 34.8 | 53.6 | 56.6 | 40.9 | 43.2 | 30.4 | 39.2 | — | |
| Gemini-1.5-Pro2026.03 | 45.4 | 56.2 | 30.9 | 64.1 | 43.6 | 51.3 | 46.3 | 36 | 34.6 | — | |
| VILASR-7B2026.03 | 45.4 | 63.5 | 34.4 | 60.6 | 30.9 | 48.9 | 45.2 | 30.4 | 49.2 | — | |
| Gemini-1.5-ProModel Category=Proprietary Models (API)2026.03 | 45.4 | 56.2 | 30.9 | 64.1 | 43.6 | 51.3 | 46.3 | 36 | 34.6 | — | |
| Gemini-1.5 ProModel Category=Proprietary, Frame Sampling Strategy=1 FPS2025.05 | 45.4 | 56.2 | 30.9 | 64.1 | 43.6 | 51.3 | 46.3 | 36 | 34.6 | — | |
| SpaceR-7B2026.03 | 43.5 | 61.9 | 28.6 | 60.9 | 35.2 | 38.2 | 46 | 31.4 | 45.6 | — | |
| SAB-LVLMModel=InternVL3.5-38B-Instruct, Weight bits=1.06 bit, Memory=10.27 GB2026.07 | 43 | — | — | — | — | — | — | — | — | — | |
| Gemini-1.5-FlashModel Category=Proprietary Models (API)2026.03 | 42.1 | 49.8 | 30.8 | 53.5 | 54.4 | 37.7 | 41 | 31.5 | 37.8 | — | |
| ARB-LLMModel=InternVL3.5-38B-Instruct, Weight bits=1.08 bit, Memory=10.27 GB2026.07 | 41.62 | — | — | — | — | — | — | — | — | — | |
| SPAR-8BModel Category=Spatial-Enhanced Models2026.03 | 41.1 | — | — | — | — | — | — | — | — | — | |
| LLaVA-NeXT-Video-72BModel Category=Open-source Models2026.03 | 40.9 | 48.9 | 22.8 | 57.4 | 35.3 | 42.4 | 36.7 | 35 | 48.6 | — | |
| LLaVA-Video-72BModel Category=Open-source, Number of Frames=8-32 frames2025.05 | 40.9 | 48.9 | 22.8 | 57.4 | 35.3 | 42.4 | 36.7 | 35 | 48.6 | — | |
| LLaVA-OneVision-72BModel Category=Open-source Models2026.03 | 40.2 | 43.5 | 23.9 | 57.6 | 37.5 | 42.5 | 39.9 | 32.5 | 44.6 | — | |
| LLaVA-OneVision-72BModel Category=Open-source, Number of Frames=8-32 frames2025.05 | 40.2 | 43.5 | 23.9 | 57.6 | 37.5 | 42.5 | 39.9 | 32.5 | 44.6 | — | |
| Qwen2.5-VL-72BModel Category=Open-source, Number of Frames=162025.05 | 37 | 25.1 | 29.3 | 57.9 | 29.4 | 41.7 | 37 | 23.2 | 29 | — | |
| Keye-VL-1.5-8BModel Category=Open-Weight Models, Parameter count=1.5-8B2026.06 | 36.4 | — | — | — | — | — | — | — | — | — | |
| InternVL2-40BModel Category=Open-source Models2026.03 | 36 | 34.9 | 26.9 | 46.5 | 31.8 | 42.1 | 32.2 | 34 | 39.6 | — | |
| InternVL2-40BModel Category=Open-source, Number of Frames=8-32 frames2025.05 | 36 | 34.9 | 26.9 | 46.5 | 31.8 | 42.1 | 32.2 | 34 | 39.6 | — | |
| LLaVA-Video-7B2026.03 | 35.6 | 48.5 | 14 | 47.8 | 24.2 | 43.5 | 42.4 | 34 | 30.6 | — | |
| LLaVA-NeXT-Video-7BModel Category=Open-source Models2026.03 | 35.6 | 48.5 | 14 | 47.8 | 24.2 | 43.5 | 42.4 | 34 | 30.6 | — | |
| InternVL2-8BModel Category=Open-source Models2026.03 | 34.6 | 23.1 | 28.7 | 48.2 | 39.8 | 36.7 | 30.7 | 29.9 | 39.6 | — | |
| GPT-4o2026.03 | 34 | 46.2 | 5.3 | 43.8 | 38.2 | 37 | 41.3 | 31.5 | 28.5 | — | |
| GPT-4oModel Category=Proprietary Models (API)2026.03 | 34 | 46.2 | 5.3 | 43.8 | 38.2 | 37 | 41.3 | 31.5 | 28.5 | — | |
| GPT-4oModel Category=Proprietary, Number of Frames=8-32 frames2025.05 | 34 | 46.2 | 5.3 | 43.8 | 38.2 | 37 | 41.3 | 31.5 | 28.5 | — | |
| InternVideo2.5-7BModel Category=Open-Weight Models, Parameter count=7B2026.06 | 33.4 | — | — | — | — | — | — | — | — | — | |
| Qwen2.5-VL-7BModel Category=Open-source, Number of Frames=162025.05 | 33 | 40.9 | 14.8 | 43.4 | 10.7 | 38.6 | 38.5 | 33 | 29.8 | — | |
| VideoChat-Flash-7BModel Category=Open-Weight Models, Parameter count=7B2026.06 | 32.8 | — | — | — | — | — | — | — | — | — | |
| Qwen2.5-VL-7B2026.03 | 32.7 | 34.5 | 19.4 | 47.6 | 40.8 | 32.8 | 24.5 | 32.5 | 29.4 | — | |
| SAB-LVLMModel=InternVL3.5-14B-Instruct, Weight bits=1.07 bit, Memory=4.33 GB2026.07 | 32.5 | — | — | — | — | — | — | — | — | — | |
| LLaVA-OneVision-7B2026.03 | 32.4 | 47.7 | 20.2 | 47.4 | 12.3 | 42.5 | 35.2 | 29.4 | 24.4 | — | |
| LLaVA-OneVision-7BModel Category=Open-source Models2026.03 | 32.4 | 47.7 | 20.2 | 47.4 | 12.3 | 42.5 | 35.2 | 29.4 | 24.4 | — | |
| MiniCPM-V-4.5-8BModel Category=Open-Weight Models, Parameter count=4.5-8B2026.06 | 32 | — | — | — | — | — | — | — | — | — | |
| ARB-LLMModel=InternVL3.5-14B-Instruct, Weight bits=1.07 bit, Memory=4.33 GB2026.07 | 31.5 | — | — | — | — | — | — | — | — | — | |
| Gemini-2.5-Pro (minimal)Size=-2026.03 | 31.4 | — | — | — | — | — | — | — | — | — | |
| VILA-1.5-40BModel Category=Open-source Models2026.03 | 31.2 | 22.4 | 24.8 | 48.7 | 22.7 | 40.5 | 25.7 | 31.5 | 32.9 | — | |
| VILA-1.5-40BModel Category=Open-source, Number of Frames=8-32 frames2025.05 | 31.2 | 22.4 | 24.8 | 48.7 | 22.7 | 40.5 | 25.7 | 31.5 | 32.9 | — | |
| Qwen2.5-VL-3BModel Category=Open-source, Number of Frames=162025.05 | 30.6 | 24.3 | 24.7 | 31.7 | 22.6 | 38.3 | 41.6 | 26.3 | 21.2 | — | |
| LongVA-7BModel Category=Open-source Models2026.03 | 29.2 | 38 | 16.6 | 38.9 | 22.2 | 33.1 | 43.3 | 25.4 | 15.7 | — | |
| LongVA-7BModel Category=Open-source, Number of Frames=8-32 frames2025.05 | 29.2 | 38 | 16.6 | 38.9 | 22.2 | 33.1 | 43.3 | 25.4 | 15.7 | — | |
| VILA-1.5-8BModel Category=Open-source Models2026.03 | 28.9 | 17.4 | 21.8 | 50.3 | 18.8 | 32.1 | 34.8 | 31 | 24.8 | — | |
| LLaVA-OneVision-0.5BModel Category=Open-source Models2026.03 | 28 | 46.1 | 28.4 | 15.4 | 28.3 | 28.9 | 36.9 | 34.5 | 5.8 | — | |
| InternVL2-2BModel Category=Open-source Models2026.03 | 27.4 | 21.8 | 24.9 | 22 | 35 | 33.8 | 44.2 | 30.5 | 7.1 | — | |
| PB-LLMModel=InternVL3.5-38B-Instruct, Weight bits=1.70 bit, Memory=10.24 GB2026.07 | 26.83 | — | — | — | — | — | — | — | — | — | |
| BiLLMModel=InternVL3.5-38B-Instruct, Weight bits=1.07 bit, Memory=10.72 GB2026.07 | 26.71 | — | — | — | — | — | — | — | — | — | |
| LongVILA-8BModel Category=Open-source Models2026.03 | 21.6 | 29.1 | 9.1 | 16.7 | 0 | 29.6 | 30.7 | 32.5 | 25.5 | — | |
| LongVILA-8BModel Category=Open-source, Number of Frames=8-32 frames2025.05 | 21.6 | 29.1 | 9.1 | 16.7 | 0 | 29.6 | 30.7 | 32.5 | 25.5 | — | |
| BiLLMModel=InternVL3.5-14B-Instruct, Weight bits=1.07 bit, Memory=4.53 GB2026.07 | 17.36 | — | — | — | — | — | — | — | — | — | |
| SAB-LVLMModel=InternVL3.5-8B-Instruct, Weight bits=1.07 bit, Memory=2.28 GB2026.07 | 9.63 | — | — | — | — | — | — | — | — | — | |
| BiLLMModel=InternVL3.5-8B-Instruct, Weight bits=1.08 bit, Memory=2.39 GB2026.07 | 1.24 | — | — | — | — | — | — | — | — | — | |
| PB-LLMModel=InternVL3.5-8B-Instruct, Weight bits=1.70 bit, Memory=2.28 GB2026.07 | 0.26 | — | — | — | — | — | — | — | — | — | |
| ARB-LLMModel=InternVL3.5-8B-Instruct, Weight bits=1.07 bit, Memory=2.28 GB2026.07 | 0.16 | — | — | — | — | — | — | — | — | — | |
| PB-LLMModel=InternVL3.5-14B-Instruct, Weight bits=1.70 bit, Memory=4.34 GB2026.07 | 0.09 | — | — | — | — | — | — | — | — | — | |
| Gemma4Parameters=26BA4B, Internal Reasoning (Think mode)=false, Input Frames=322026.05 | — | — | — | — | — | — | — | — | — | 32.91 | |
| Qwen3.5Parameters=9B, Internal Reasoning (Think mode)=false, Input Frames=1282026.05 | — | — | — | — | — | — | — | — | — | 55.67 | |
| Qwen3.5Parameters=35BA3B, Internal Reasoning (Think mode)=false, Input Frames=1282026.05 | — | — | — | — | — | — | — | — | — | 58.1 | |
| Qwen3VLParameters=8B, Internal Reasoning (Think mode)=true, Input Frames=1282026.05 | — | — | — | — | — | — | — | — | — | 56.61 | |
| Qwen3VLParameters=30BA3B, Internal Reasoning (Think mode)=true, Input Frames=1282026.05 | — | — | — | — | — | — | — | — | — | 51.56 | |
| Random2026.03 | — | — | — | — | — | 25 | 36.1 | 28.3 | 25 | — | |
| SAT-LLaVA-Video-7BModel Category=Spatial-Enhanced Models2026.03 | — | — | — | — | 47.3 | 41.1 | 37.1 | 36.1 | 40.4 | — | |
| SenseNova-U1Parameters=8B, Internal Reasoning (Think mode)=true, Input Frames=322026.05 | — | — | — | — | — | — | — | — | — | 62.66 | |
| SenseNova-U1Parameters=30BA3B, Internal Reasoning (Think mode)=true, Input Frames=322026.05 | — | — | — | — | — | — | — | — | — | 56.9 |