> ## Documentation Index
> Fetch the complete documentation index at: https://mintlify.com/NVIDIA/Isaac-GR00T/llms.txt
> Use this file to discover all available pages before exploring further.
# Introduction
> NVIDIA Isaac GR00T N1.6 is an open vision-language-action model for generalized humanoid robot skills
## What is GR00T N1.6?
NVIDIA Isaac GR00T N1.6 is an open vision-language-action (VLA) model for generalized humanoid robot skills. This cross-embodiment model takes multimodal input, including language and images, to perform manipulation tasks in diverse environments.
GR00T N1.6 is trained on a diverse mixture of robot data including bimanual, semi-humanoid and an expansive humanoid dataset. It is adaptable through post-training for specific embodiments, tasks and environments.
GR00T N1.6 represents a significant upgrade over GR00T N1.5, with improvements in both model architecture and data leading to better performance across benchmarks.
## Key capabilities
Trained on 10,000+ hours of robot data from diverse embodiments including bimanual arms, semi-humanoid robots, and full humanoid platforms like Unitree G1 and Galaxea R1 Pro.
Processes vision, language, and proprioceptive state inputs using a 2B parameter vision-language backbone with flexible resolution support.
Generates smooth, continuous actions through a 32-layer diffusion transformer that denoises action trajectories.
Predicts state-relative action chunks for most embodiments, improving generalization across different robot configurations.
Achieves 27.3 Hz on RTX 5090 with torch.compile, with even faster TensorRT deployment options available.
Fine-tune on small custom datasets to adapt the foundation model to specific robotics tasks with minimal data.
## Model architecture
The neural network architecture of GR00T N1.6 combines a vision-language foundation model with a diffusion transformer head that denoises continuous actions:
The architecture consists of three main components:
1. **Vision-language backbone**: NVIDIA Cosmos-Reason-2B VLM variant with flexible resolution and native aspect ratio support
2. **Action head**: 32-layer diffusion transformer (DiT) with cross-attention to VLM features
3. **Embodiment-specific projectors**: Category-specific MLPs for encoding states and decoding actions per robot
## Performance benchmarks
GR00T N1.6 achieves state-of-the-art results across multiple simulation benchmarks:
| Benchmark | Task Type | Success Rate |
| -------------- | --------------------- | ----------------- |
| LIBERO-Spatial | Tabletop manipulation | High performance |
| SimplerEnv | Bimanual tasks | Competitive |
| BEHAVIOR-1K | Loco-manipulation | Strong results |
| RoboCasa | Kitchen tasks | Zero-shot capable |
Inference timing on RTX 5090 with 4 denoising steps: 37ms end-to-end (18ms backbone + 16ms action head), achieving 27.3 Hz throughput.
## Getting started
Set up your environment with uv package manager and install dependencies
Run inference with pre-trained checkpoints in minutes
Adapt GR00T to your robot embodiment and tasks
Explore simulation benchmarks and deployment examples
## Target audience
GR00T N1.6 is intended for researchers and professionals in robotics. This repository provides tools to:
* Leverage a pre-trained foundation model for robot control
* Fine-tune on small, custom datasets
* Adapt the model to specific robotics tasks with minimal data
* Deploy the model for inference on real hardware
The focus is on enabling customization of robot behaviors through fine-tuning.
## Resources
Read the full technical paper on arXiv
Download pre-trained checkpoints from Hugging Face
Explore the official research blog post
Access the cross-embodiment training dataset