RoboTrustBench: Benchmarking the Trustworthiness of Video World Models for Robotic Manipulation
For robotic manipulation researchers, this benchmark reveals critical trustworthiness gaps in video world models beyond visual quality.
RoboTrustBench evaluates video world models for robotic manipulation under normal, constraint-sensitive, counterfactual, and adversarial scenarios. Current models generate visually coherent videos but fail at constraint reasoning, counterfactual grounding, and unsafe-instruction suppression.
Video world models are increasingly used in robotic manipulation, yet existing benchmarks mostly evaluate them under valid, feasible, and safe instructions. We introduce RoboTrustBench, a benchmark for evaluating the trustworthiness of video world models under four scenarios: Normal, Constraint-Sensitive, Counterfactual, and Adversarial. Built from real-world DROID episodes, RoboTrustBench contains 1,207 expert-validated instruction-image pairs and a six-dimensional evaluation protocol with 13 fine-grained criteria. Evaluating seven representative video world models with human and MLLM assessment, we find that current models often generate visually coherent videos, but struggle with constraint reasoning, counterfactual grounding, physical interaction, and unsafe-instruction suppression. These results show that visual quality and surface-level instruction following are insufficient for trustworthy robotic video world modeling.