Generalist publishes research preview of autonomous dexterous manipulation system
Evidence notes
- Generalist shared its first public research preview showing fully autonomous end-to-end models controlling robots in real time across dexterous manipulation tasks.
- The company highlighted cross-embodiment transfer across different robot arms, zero task data in some evaluated environments, and 100 Hz control from pixels and sensor inputs.
- The preview established the company's initial public technical baseline ahead of later GEN-series model releases.
Company context
Generalist builds embodied foundation models for robots, initially focused on dexterous manipulation. Its work combines real-world robot interaction data, in-house data-collection hardware and models intended to generalize across tasks.