BiDex-100 tests robot policies across 106 coordinated two-handed tasks

The simulation benchmark evaluates dexterous vision-language-action models by capability, distribution shift and data efficiency rather than success alone.

Chinese Tech
Yong Wang · Kecheng Zheng · Xing Zhu · Chao Xu · Yujun Shen · Shengze Cai

Zhejiang University · Ant Group

Research Digest··2 min read
Wang and colleagues introduce BiDex-100, a MuJoCo benchmark for controlling two multi-fingered robotic hands across 106 manipulation tasks.

The authors built tasks for a 58-degree-of-freedom bimanual system, a substantially larger action space than conventional robot arms with parallel-jaw grippers.

§

Research Digest

Written by Alwyn Sage, research editor. Plate by Ollie Flash, picture desk.

first published
Meet the newsroomHow we workSubscribe