cs.RONov 19, 2026

PhyDEX: A Benchmark for Physical Reasoning in Vision-Language-Action Models via Dynamic Extrinsic Dexterity

Authors: Hao Vo, Taisei Hanyu, Yuki Ikebe, Pham Tri Quang, Sreevenkata Anjani Tishita Godavarthi, Nghi D. Q. Bui, Daniel Sonntag, Chase Rainwater, 

Organizations: AICV Lab, University of Arkansas, USA · VinUniversity, Vietnam · Center for AI Research, VinUniversity, Vietnam · University of Stuttgart, Germany · 

Abstract

Vision-Language-Action (VLA) models have advanced rapidly in language-conditioned robotic manipulation, yet existing benchmarks mainly evaluate gripper--object contact and largely ignore how objects interact with the environment through gravity, friction, inertia, and mass distribution. As a result, it remains unclear whether current policies truly understand physics or merely interpolate from their training data. We address this gap through the lens of dynamic extrinsic dexterity and introduce PhyDEX, a throwing benchmark in which a robot must infer an object's mass, center of mass, and friction to send it to a destination beyond its kinematic workspace. PhyDEX contains 35K trajectories (143 hours) across nine tasks and 100 scenes, covering uniform, irregular, and dynamic-mass objects with static, multiple, and moving destinations. All data are generated by a scalable, fully automatic pipeline that combines grasp generation, motion planning, and closed-loop throw optimization under randomized physical properties. We evaluate three representative VLAs, GR00T N1.7, DynamicVLA, and ForceVLA, and find that all achieve success rates below 3%. Our analysis reveals two key bottlenecks: limited robustness of pretrained models, which causes most failures before the throw, and physical blindness, as providing ground-truth physical properties roughly doubles success. Experiments on a humanoid robot and in the real world show the same failure patterns. We hope PhyDEX will drive the development of physically grounded robot policies.

Figures & tables

Explore similar work

CardsList
  1. PhysVLA: Towards Physically-Grounded VLA for Embodied Robotic Manipulation

    Jun 11, 2026Namai Chandra, Shriram Damodaran, Lin WangRobotic ControlRobotic Manipulation

  2. Position: Vision-Language-Action Models Cannot Be Verified to Perform Physical Reasoning

    Jun 28, 2026Taozhao Chen, Ian Manchester, Huaming ChenVLM EvaluationPhysical Reasoning

  3. Rethinking the Practicality of Vision-language-action Model: A Comprehensive Benchmark and An Improved Baseline

    Feb 26, 2026Wenxuan Song, Jiayi Chen, Xiaoquan Sun +12Vision-Language-Action ModelsEfficient VLAs