●builderThis benchmark provides a rigorous testing ground for geospatial AI applications requiring logical spatial reasoning.
●researcherYou can now evaluate models on their ability to resolve complex, relational spatial queries rather than just segmentation accuracy.