GAUGE
GAUGE evaluates physical fidelity of simulation engines and video world models using 22 task families covering rigid bodies, cables, textiles, and deformable objects, with real-world trajectories and calibrated metadata.
- Released
- 2026-08-06
- Readiness
- Paper only
- Primary field
- Robotics & Autonomous Systems
Why it matters
Existing evaluations of physical fidelity rely on perceptual similarity or human judgment. GAUGE provides a diagnostic benchmark with grounded physical measurements, enabling systematic comparison of simulators and world models on specific physical principles.
Motivation
Physics engines facilitate large-scale training and evaluation for embodied intelligence, while generative video world models are emerging as implicit simulators of future states and interactions.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.