Build physical AI faster with NVIDIA Cosmos: The advanced world foundation model platform.   [Get Started](/cosmos)

1. 
  

[Home](https://developer.nvidia.com/)2. 
3. [Isaac](/isaac)
4. 
Isaac GR00T

# NVIDIA Isaac GR00T

## Generalist Robot 00 Technology  

NVIDIA Isaac™ GR00T is a research initiative and development platform for developing general-purpose robot foundation models and data pipelines to accelerate humanoid robotics research and development.

[Download Isaac GR00T N Models](https://github.com/NVIDIA/Isaac-GR00T)[Experience the GR00T-Mimic Blueprint](https://build.nvidia.com/nvidia/isaac-gr00t-synthetic-manipulation)

* * *

## See Isaac GR00T in Action  

https://www.youtube-nocookie.com/embed/Coy2TyBcT4g?
 

* * *

## How Isaac GR00T Works

Isaac GR00T features robot foundation models for cognition and control, simulation frameworks built on [NVIDIA Omniverse](https://developer.nvidia.com/omniverse)™ and [Cosmos](https://developer.nvidia.com/cosmos)™, data pipelines for generating synthetic data and environments, and a computer in the robot—Jetson AGX Thor™—to run the entire robot stack.

 ![A diagram showing how NVIDIA Isaac Gr00T works](https://developer.download.nvidia.com/images/products/gr00t-1920-1080.jpg)

### Introductory Video   

Watch how Isaac GR00T helps provide the building blocks for the future of AI-powered robotics.

[Watch Video (2:21)](https://www.youtube.com/watch?v=m1CH-mgpdYg)

### Introductory Blog  

Explore the world’s first open foundation model for generalized humanoid robot reasoning and skills.

[Read Blog](https://developer.nvidia.com/blog/accelerate-generalist-humanoid-robot-development-with-nvidia-isaac-gr00t-n1/)

### 
### Train Humanoids With Isaac GR00T-Dreams

Watch how GR00T-Dreams and Isaac GR00T are enabling robots to learn new tasks and generalize across environments.

[Watch Video](https://www.youtube.com/watch?v=pMWL1MEI-gE)

### Developer Tutorial 

Learn how to post-train Isaac GR00T N1.5, enabling robot training with just 20-40 demonstrations.

[Watch Tutorial](https://www.youtube.com/watch?v=L-WriIVOyYE&t=4s)

* * *

## Isaac GR00T N Robot Foundation Models

Isaac GR00T N open foundation models are ideal for generalized humanoid robot reasoning and skills. This cross-embodiment solution takes multimodal input—including language and images—to perform manipulation tasks in diverse environments.   
  
These models are trained on an expansive humanoid dataset consisting of real captured data, synthetic data generated using the components of the [Isaac GR00T-Mimic blueprint](https://build.nvidia.com/nvidia/isaac-gr00t-synthetic-manipulation), [Isaac GR00T-Dreams blueprint](http://github.com/nvidia/gr00t-dreams), and internet-scale video data. They’re also adaptable through post-training for specific embodiments, tasks, and environments.  
  
Isaac GR00T N models can easily generalize across common tasks—such as grasping, moving objects with one or both arms, and transferring items from one arm to another—or perform multi-step tasks that require long context and combinations of general skills. These capabilities can be applied across a variety of use-cases, including material handling, packaging, and inspection.

### Isaac GR00T N1.7 3B Model

[Download From Hugging Face](https://huggingface.co/nvidia/GR00T-N1.7-3B)

* * *

## Isaac GR00T Workflows

Discover powerful tools for accelerating the development of advanced capabilities for humanoid robots and bridge the simulation-to-reality gap.

### Isaac-Teleop

Collect high-quality human demonstrations through teleoperation in the real world and simulation.

[Learn More About Isaac-Teleop](https://github.com/NVIDIA/IsaacTeleop)

 ![A diagram showing how NVIDIA Isaac Gr00T works](https://developer.download.nvidia.com/icons/m48-robot-mimic-r4.svg)
### GR00T-Mimic

Generate a large number of synthetic motion trajectories from a small number of human demonstrations.

[Learn More About the GR00T-Mimic Blueprint](https://build.nvidia.com/nvidia/isaac-gr00t-synthetic-manipulation)

 ![Placeholder](https://developer.download.nvidia.com/icons/m48-robot-reinforcement-learning-alt-256px-blk.png)
### GR00T-Gen

Teach robots to generalize and adapt to new situations using diverse, realistic simulation environments.

[Learn More About GR00T-Gen](https://developer.nvidia.com/blog/advancing-humanoid-robot-sight-and-skill-development-with-nvidia-project-gr00t/#gr00t-gen_for_diverse_environment_generation)

 ![Placeholder](https://developer.download.nvidia.com/icons/m48-robot-hand-256px-blk.png)
### GR00T-Dexterity

Simplify the complex grasping tasks with an end-to-end, pixels- to- action grasping system to have human-like dexterity.

[Learn More About GR00T-Dexterity](https://developer.nvidia.com/blog/advancing-humanoid-robot-sight-and-skill-development-with-nvidia-project-gr00t/#gr00t-dexterity_for_fine-grained_and_dexterous_manipulation%C2%A0)

 ![Placeholder](https://developer.download.nvidia.com/icons/m48-foundational-model-dexterity-256px-blk.png)
### GR00T-Mobility  

Enhance the robot’s ability to adapt to new, unseen environments using reinforcement learning and imitation learning techniques.

[Learn More About GR00T-Mobility](https://developer.nvidia.com/blog/advancing-humanoid-robot-sight-and-skill-development-with-nvidia-project-gr00t/#gr00t-mobility_for_locomotion_and_navigation%C2%A0)

 ![Placeholder](https://developer.download.nvidia.com/icons/m48-robot-humanoid-256px-blk.png)
### GR00T-Control  

Achieve responsive and precise humanoid robot control with a suite of whole- body control libraries, models, and policies.

[Learn More About GR00T-Control](https://developer.nvidia.com/blog/advancing-humanoid-robot-sight-and-skill-development-with-nvidia-project-gr00t/#gr00t-control_for_whole-body_control)

 ![Placeholder](https://developer.download.nvidia.com/icons/m48-foundational-model-cognition-256px-blk.png)
### GR00T-Perception  

Integrate vision language models, LLMs, and retrieval-augmented memory to boost perception, cognition, and adaptability.

[Learn More About GR00T-Perception](https://developer.nvidia.com/blog/advancing-humanoid-robot-sight-and-skill-development-with-nvidia-project-gr00t/#gr00t-perception_for_multimodal_sensing)

* * *

## Get Started Developing Humanoid Robots

Get the software and infrastructure you need to advance your humanoid robot development with GR00T foundational technologies.

### Software and Services

#### Data Generation Pipelines

Generate large synthetic trajectory data sets from just a few human demonstrations with GR00T-Mimic and Cosmos. Then, train your robot motion policy model in Isaac Lab.

[Try It Now](https://github.com/NVIDIA-Omniverse-blueprints/synthetic-motion-generation)
####   

Generate vast amounts of synthetic trajectory data using NVIDIA Cosmos world foundation models to teach robots how to perform new actions in novel environments.

[Try It Now](http://github.com/nvidia/gr00t-dreams)

#### Train Robot Policies  

NVIDIA Isaac Lab enables scalable, adaptable policy training in physically accurate scenes, bridging the sim-to-real gap.

[Learn More About Isaac Lab](https://developer.nvidia.com/isaac/lab)

#### Simulate and Validate   

The trained robot policies can be validated in physically accurate environments using NVIDIA Isaac Sim™ before deployment. 

[Learn More About Isaac Sim](https://developer.nvidia.com/isaac/sim)

### Compute Infrastructure

#### Train  

The NVIDIA DGX™ Cloud end-to-end AI platform for training robotics foundation models gives you scalable capacity built on the latest NVIDIA architecture. It’s co-engineered with the world’s leading cloud service providers.

[Try DGX Now](https://www.nvidia.com/en-us/data-center/get-dgx/)

#### Simulate  

NVIDIA Omniverse and Cosmos systems deliver the simulation platforms on industry-leading NVIDIA RTX™ PRO 6000 Blackwell Workstation and Server GPUs. This lets you accelerate the next generation of robotics simulation and learning workloads.

[Explore the NVIDIA RTX PRO 6000 Blackwell Series  
](https://www.nvidia.com/en-us/products/workstations/professional-desktop-gpus/rtx-pro-6000-family/)

#### Deploy  

Accelerate the development of advanced humanoid robots and run multimodal AI models. Jetson AGX Thor is based on NVIDIA Blackwell architecture and comes with integrated functional safety, high-performance CPU, and 100 GB of ethernet bandwidth.

[Learn More  
](https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/)

* * *

## Humanoid Developer Learning Library

[

Whitepaper 

### An Open Foundation Model for Generalist Humanoid Robots

A robot foundation model, trained on massive and diverse data sources, is essential for enabling the robots to reason about novel situations, robustly handle real-world variability, and rapidly learn new tasks. .

](https://arxiv.org/abs/2503.14734)

[

Tech Blog 

### Building a Synthetic Motion Generation Pipeline for Humanoid Robot Learning

Learn how imitation learning enables developers to generate exponentially large datasets from a small number of human demonstrations.  
.

](https://developer.nvidia.com/blog/building-a-synthetic-motion-generation-pipeline-for-humanoid-robot-learning/)

[

Tech Blog 

### Advancing Physical AI with NVIDIA Cosmos World Foundation Model Platform

NVIDIA Cosmos is a platform that helps developers build custom world models for physical AI systems at scale. It offers open world foundation models and tools for every stage of development, from data curation to training to customization.

](https://developer.nvidia.com/blog/advancing-physical-ai-with-nvidia-cosmos-world-foundation-model-platform/)

* * *

## Humanoid Robotics Ecosystem

NVIDIA builds foundation models, acceleration libraries, and blueprints to accelerate the world’s ecosystem of humanoid developers.

 ![NVIDIA’s ecosystem of humanoid developers - Ability Robotics](https://developer.download.nvidia.com/images/logos/agility-robotics-logo.svg)

 ![NVIDIA’s ecosystem of humanoid developers - Apptronik](https://developer.download.nvidia.com/images/logos/logo-apptronik.svg)

 ![NVIDIA’s ecosystem of humanoid developers - Boston Dynamics](https://developer.download.nvidia.com/images/logos/boston-dynamics-logo.svg)

 ![NVIDIA’s ecosystem of humanoid developers - Neura Robotics](https://developer.download.nvidia.com/images/logos/neura-robotics-logo-black.svg)

 ![NVIDIA’s ecosystem of humanoid developers - Sanctuary AI](https://developer.download.nvidia.com/images/logos/sanctuary-ai-logo.svg)

 ![NVIDIA’s ecosystem of humanoid developers - Unitree Robotics](https://developer.download.nvidia.com/images/logos/unitree-logo.svg)

* * *

## More Resources

 ![A decorative image representing Developer Community](https://developer.download.nvidia.com/icons/m48-people-group.svg)

### Explore the Community  

 ![img-alt-text](https://developer.download.nvidia.com/icons/m48-certification-ribbon-2.svg)

### Get Training and Certification  

 ![A decorative image representing Inception for Startups](https://developer.download.nvidia.com/icons/m48-ai-startup.svg)

### Join the Program for Startups  

## Accelerate Your Humanoid Robot Development  

[Download Isaac GR00T N Models](https://github.com/NVIDIA/Isaac-GR00T?)[Get Started with Isaac Lab](https://developer.nvidia.com/isaac/lab)

## FAQs

GR00T N models accept three multimodal inputs: (1) a video sequence from onboard cameras, (2) a natural language command, and (3) the robot’s current proprioceptive state (joint positions).

The model outputs action chunks—predictive sequences of relative joint motions. By predicting multiple steps into the future, the model ensures smooth whole-body control, enabling stable locomotion and coordinated bimanual manipulation.

GR00T N models utilize a massive mixture of data sources: internet-scale human video for world knowledge, real-world teleoperation for physical grounding, and synthetic data generated to scale learning safely.

  

This diversity enables strong generalization across different hardware. For example, the GR00T N1.6 training set expands embodiment coverage to include the Unitree G1, AgiBot Genie-1, and Fourier GR-1, alongside specialized bimanual manipulation datasets (YAM).

GR00T N models can be adapted to new hardware by fine-tuning them with a small collection of recorded task demonstrations. This efficient process allows the system to learn specific robot movements while preserving the broad general capabilities established during pre-training.

Isaac GR00T N1.7 Early Access allows developers to begin experimenting with the model and integrating it into their workflows ahead of the full commercial release. Developers can download the model, explore the codebase, and begin building production-oriented applications on the NVIDIA Isaac GR00T stack, with the understanding that support and stability guarantees will be limited until the full commercial release. 

What is supported in Early Access

- Access to the pre-trained GR00T N1.7 model weights and reference code.

- The ability to fine-tune or run inference with the model using custom robot data or demonstrations.

- Experimentation, prototyping, and research use cases.

What is not supported

- Production deployment with commercial support.

- A stable feature set with thoroughly validated performance.

- Product-level support from NVIDIA

[Isaac Lab](https://developer.nvidia.com/isaac/lab) provides an open, modular framework to train robot policies using reinforcement learning and imitation learning. Isaac Lab is natively integrated with Isaac Sim for high-fidelity physics and rendering, Newton for an extensible physics engine, and can also use Isaac GR00T N open models and blueprints to accelerate skill learning across different robot embodiments.