π*0.6
$ 0
π*0.6 is Physical Intelligence’s robot foundation model that learns from real-world practice. It improves through human corrections and its own experience, which makes everyday tasks faster and more reliable.
Available on backorder
Brain Score
4
Specifications and details:
| Nationality | US |
|---|---|
| Website | https://www.pi.website/ |
| Model type | Foundation Model |
| Manufacturer | Physical Intelligence |
| Release date | November 17, 2025 |
Description
π*0.6 is a robot foundation model from Physical Intelligence, a robotics AI startup based in San Francisco. The company introduced it in November 2025. It builds on π0.6, the company’s base model that connects vision, language and action. However, π*0.6 adds something new. It keeps improving after training by learning from its own experience on real robots. As a result, it moves past the limits of simply copying human demonstrations.
Physical Intelligence calls its training method Recap. First, people show the robot how to do a task. Next, they step in to correct its mistakes. Finally, the robot practices on its own and learns which actions lead to success. This approach more than doubled the robot’s speed on some of the hardest jobs. For example, π*0.6 made espresso drinks from early morning until late at night. It also folded laundry in a new home and assembled packaging boxes in a real factory.
New! 2026 Humanoid
Robot Market Report
198 pages of exclusive insight from global robotics experts — uncover funding trends, technology challenges, leading manufacturers, supply chain shifts, and surveys and forecasts on future humanoid applications.
now Google DeepMind
Contact Humanoid.guide
Website: https://www.pi.website/blog/pistar06
