A compact 450M-parameter VLA from Hugging Face, pretrained on community-collected LeRobot datasets and sized for fine-tuning and inference on consumer GPUs.
nvidia/GR00T-N1.6-3B
Indexed from Hugging FaceLinked to its home on Hugging Face. FrontierOS indexes it; nothing is hosted or mirrored here.ModelVision-language-action3.3B paramsEst. Edge GPUEstimated minimum compute for inference: Jetson Orin class, 8–16 GB shared memory. Not a benchmark.NVIDIA One-Way Noncommercial LicenseCustom or unstated terms. The upstream license governs use.HumanoidBimanualSingle arm
About
A 3B cross-embodiment robot foundation model from NVIDIA that combines a vision-language backbone with a diffusion-transformer action head and can be post-trained on a new robot with a modest amount of data.
Why it matters for physical AI
It is one of the main open-weight policies aimed at humanoids and whole-body manipulation, though its noncommercial license limits product use.
Summary written by FrontierOS in our own words. The upstream card and license govern use. Metadata checked 27 Sep 2026. Read the full card on Hugging Face
Files, weights and version history live on Hugging Face. FrontierOS indexes this repo; it does not host or mirror it.
hf download nvidia/GR00T-N1.6-3BUses the Hugging Face CLI (pip install -U huggingface_hub). This downloads the whole repo; add --include "<pattern>" to fetch only the files you need.
Open humanoid data is still scarce. What are you training on?
The permissive options are few: Humanoid Everyday for real teleoperation, NVIDIA's GR00T simulation sets, and large egocentric human video as a proxy.
More vision-language-action
Browse allLeRobot's port of the Physical Intelligence pi0.5 base model, which is co-trained on web and multi-robot data to generalize better to unfamiliar homes and objects.
A 7B vision-language-action model on a Llama 2 backbone that maps one camera image and a text instruction to discretized 7-DoF end-effector actions, trained on Open X-Embodiment episodes.