RT-2Google DeepMind
Robot foundation modelsClosed
Google DeepMind · AI model
RT-2
July 2023 model that popularised the term vision-language-action. It took large vision-language models (PaLI-X and PaLM-E) and trained them to output robot actions as well as text, so web knowledge could carry over to robot control. Robot data came from the RT-1 fleet.
- Kind
- Robot foundation models
- Maker
- Google DeepMindUnited States
- First year
- 2023
- Weights
- ClosedOnly the maker runs it
- Robots
- 1Named as using it
- Country
- United States
Robots that use it
1 robot model named by the maker or reputable press.
Not yet in the atlas
- Everyday Robots mobile manipulators (RT-1 data fleet)deepmind.google
Other robot foundation models
Newest first.
- 2026Isaac GR00T N1.7NVIDIA · 2 robot models
- 2026π0.7Physical Intelligence · 1 robot model
- 2026Helix 02 (and Helix 2.5)Figure AI · 1 robot model
- 2026Gemini Robotics 2 (with ER 2 and On-Device 2)Google DeepMind · 1 robot model
- 2025Isaac GR00T N1NVIDIA · 1 robot model
- 2025Isaac GR00T N1.5NVIDIA · 1 robot model
- 2025Isaac GR00T N1.6NVIDIA
- 2025π0.5Physical Intelligence · 1 robot model
- 2025π*0.6 (pi-star 0.6)Physical Intelligence
- 2025HelixFigure AI · 1 robot model
More from Google DeepMind
Sources
From the maker, its papers and reputable press.
Statistics only, not advice. No affiliation, sponsorship or endorsement. How we work · Terms