Benchmark RadarAI BENCHMARK PROFILE
EarthVerse
Evaluates scientific agents on 405 reproducible tasks across 199 documented natural hazard events, scoring fine-grained answer units and task-specific rubrics.
- Released
- 2026-08-24
- Readiness
- Paper only
- Primary field
- General AI
Why it matters
Provides a rigorous, provenance-focused measure of end-to-end scientific reliability for agents dealing with dynamic Earth systems and natural hazards.
Motivation
Earth-system analysis reconstructs changing physical processes from observations that differ in source, scale, timing, and modality.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.