Robot AI model
π*0.6
π*0.6 applies RECAP reinforcement learning to a generalist VLA using demonstrations, autonomous rollouts and teleoperated corrections. The paper evaluates iterative improvement on a static bimanual robot for laundry folding, box assembly and espresso preparation.
Records last verified Oct 13, 2026 · Send a correction
Also known as pi-star-0.6, π⁎0.6
- Organization
- Physical Intelligence
- Type
- VLA
- Released / introduced
- Nov 18, 2025
- Public weights
- Not established
- Licence / scope
- Not established
- Parameters
- Not disclosed
- Base model
- Gemma 3 4B (π0.6 architecture)
- Humanoid association
- None reported
- Inputs
- vision, language, proprioception, advantage condition
- Outputs
- continuous joint-position and gripper action chunks, subtask text
Release dates can mark an announcement or paper introduction. Public checkpoints, code licences and weight terms are distinct; follow the publisher’s links for the stated scope.
Documented embodiments
Source-associated hardware; this list does not establish plug-and-play compatibility or commercial robot capability.
- Static bimanual system with two 6-DoF arms
Parallel-jaw grippers; exact commercial arm model is not identified.
Disclosed training data
- Multi-robot demonstrations, web vision-language data, autonomous rollouts and expert interventions
More robot AI models
Evidence
Sources
- [s1] π*0.6: a VLA That Learns From Experience ↗
Physical Intelligence · paper · published Nov 18, 2025 · accessed Oct 13, 2026
- [s2] π*0.6 paper: Gemma 3 backbone, RECAP training and robot setup ↗
Physical Intelligence · paper · published Nov 19, 2025 · accessed Oct 13, 2026