Introducing GR00T 2: an open System One model for robotics running on onboard compute
The post announces the release of GR00T 2, described as the most advanced open System One model for robotics. It claims to run entirely on a robot's onboard compute without tethering and allows ownership of all model weights. Key features include handling long-horizon tasks like chemistry experiments or airplane assembly over 20+ steps and 5+ minutes, adapting from human wearable and simulation data via X-Adapt for zero-shot deployment on diverse robots, and unified cognition with frontier agents for physical reasoning and safety. The release includes model weights, adaptation recipes (coming December), an agentic framework, simulation engine, annotation engine, hardware design, and training data totaling 100K hours.
The announcement introduces GR00T 2, an open-source robotics model claiming to support long-horizon tasks over 20+ steps and 5+ minutes, with weights and training data (stated as 100K hours) available today while adaptation recipes are scheduled for December.
Source: Jim Fan — X Robotics · Read original article ↗
Article text · Original source · English
Introducing GR00T 2: our most advanced, open System One model for robotics. It runs entirely on your robot’s onboard compute, fully untethered, and you own every weight of it. GR00T 2 features:
- Long-horizon tasks: dexterous manipulation and whole-body control chained over 20+ steps and 5+ minutes, such as performing chemistry lab experiments or assembling model airplanes end to end.
- X-Adapt: adapting from easy data sources (human wearables and simulation) and deploying zero-shot to diverse real robots.
- Unified Cognition: native integrations with frontier agents for advanced physical reasoning, robust error recovery, and safety monitoring.
It's an open-source buffet! We are releasing not only the model weights and X-Adapt recipes (coming December), but also our agentic framework, simulation engine, annotation engine, hardware design, and even training data (100K hours, that's ~11 yrs of experience), all available today!
Everything that moves will be autonomous. Anyone can become a roboticist.
Lock in, build with us!
Deep dive in thread:
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- Control
- Reported autonomousOpen source S7
Reported numbers
training data hours
100,000 hours
View original evidence
It's an open-source buffet! We are releasing not only the model weights and X-Adapt recipes (coming December), but also our agentic framework, simulation engine, annotation engine, hardware design, and even training data (100K hours, that's ~11 yrs of experience), all available today!
Open source S6Reported duration
>5 minutes
View original evidence
1. Long-horizon tasks: dexterous manipulation and whole-body control chained over 20+ steps and 5+ minutes, such as performing chemistry lab experiments or assembling model airplanes end to end.
Open source S3training data years equivalent
≈11
View original evidence
It's an open-source buffet! We are releasing not only the model weights and X-Adapt recipes (coming December), but also our agentic framework, simulation engine, annotation engine, hardware design, and even training data (100K hours, that's ~11 yrs of experience), all available today!
Open source S6
- paper: not_reported
- weights: Reported pending Open source S6
- code: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
Everything that moves will be autonomous. Anyone can become a roboticist.
Open source S7
It's an open-source buffet! We are releasing not only the model weights and X-Adapt recipes (coming December), but also our agentic framework, simulation engine, annotation engine, hardware design, and even training data (100K hours, that's ~11 yrs of experience), all available today!
Open source S6
1. Long-horizon tasks: dexterous manipulation and whole-body control chained over 20+ steps and 5+ minutes, such as performing chemistry lab experiments or assembling model airplanes end to end.
Open source S3
Source：Jim Fan · x.com