Spatial Reasoning on SPBench (test)
70.2SI ScoreSpatialLadder
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SpatialLadderParadigm=Training-based Spatial VLMs2025.11 | 70.2 | 70.9 | 70.3 | |
| GCAParadigm=Tool-Integrated Spatial Agents2025.11 | 61.7 | 61.9 | 61.8 | |
| Gemini-2.5-ProParadigm=Baseline Foundation VLMs2025.11 | 55.6 | 58.3 | 56.3 | |
| VLaserParadigm=Training-based Spatial VLMs2025.11 | 53.2 | 69.2 | 56.9 | |
| Spatial-MLLMParadigm=Training-based Spatial VLMs2025.11 | 52 | 52 | 52 | |
| Qwen3-VL-ThinkingParadigm=Baseline Foundation VLMs2025.11 | 51.9 | 61.2 | 54.1 | |
| VILASRParadigm=Training-based Spatial VLMs2025.11 | 50.2 | 57.6 | 51.9 | |
| GLM-4.5VParadigm=Baseline Foundation VLMs2025.11 | 50 | 55.1 | 51.3 | |
| RoboBrain-2.0Paradigm=Training-based Spatial VLMs2025.11 | 49.1 | 46.8 | 48.5 | |
| TIGeRParadigm=Tool-Integrated Spatial Agents2025.11 | 48.7 | 38.8 | 46.3 | |
| SpaceRParadigm=Training-based Spatial VLMs2025.11 | 48.6 | 59.4 | 51.1 | |
| Video-R1Paradigm=Training-based Spatial VLMs2025.11 | 44.8 | 40.7 | 43.8 | |
| GPT-4oParadigm=Baseline Foundation VLMs2025.11 | 42.4 | 48.3 | 43.8 | |
| SpatialLLMParadigm=Training-based Spatial VLMs2025.11 | 32.2 | 26.4 | 30.7 |