Spatial Reasoning on QVS-Bench
74Relative DistanceHuman
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Human2025.11 | 74 | 56.12 | 59.2 | 62 | 63 | 62.96 | |
| gemini 2.5 flashModel Category=Closed-source Models2025.11 | 71 | 37.4 | 19.65 | 35 | 29 | 40.01 | |
| GPT-5Model Category=Closed-source Models2025.11 | 68 | 46.52 | 28.74 | 36 | 26 | 43.57 | |
| V2LO-7BModel Category=Qwen2.5-VL-7B Based Spatial Models2025.11 | 68 | 54.2 | 31.55 | 73 | 72 | 56.56 | |
| gemini 2.0 flashModel Category=Closed-source Models2025.11 | 61 | 50.11 | 23.4 | 27 | 26 | 40.25 | |
| GPT-4oModel Category=Closed-source Models2025.11 | 57 | 42.5 | 24.51 | 22 | 24 | 36.75 | |
| InternVL2.5-8BModel Category=Open-source Models2025.11 | 49 | 40.85 | 23.4 | 24 | 37 | 35.94 | |
| SpaceR-7BModel Category=Qwen2.5-VL-7B Based Spatial Models2025.11 | 49 | 62.25 | 24.1 | 23 | 27 | 40.09 | |
| Qwen2.5VL-7BModel Category=Qwen2.5-VL-7B Based Spatial Models2025.11 | 45 | 46.75 | 14.85 | 25 | 27 | 33.15 | |
| LLaVA-OneVision-7BModel Category=Open-source Models2025.11 | 43 | 43.12 | 18.8 | 23 | 26 | 32.36 |