OVEarth-Bench: Evaluating Category Breadth and Query Diversity for Open-Vocabulary Earth Observation

Researchers released OVEarth-Bench, a new zero-shot benchmark for open-vocabulary earth observation.

It expands evaluation to broader hierarchical categories and multiple natural-language query types.

Tests show overall performance is still limited, with multimodal large language model-based methods doing best.

EO-specific models underperform compared with the top general-purpose approaches, revealing major gaps.