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/Cosmos-Predict2.5-2B
Indexed from Hugging FaceLinked to its home on Hugging Face. FrontierOS indexes it; nothing is hosted or mirrored here.ModelWorld model2B paramsEst. DatacenterEstimated minimum compute for inference: Server-class GPU with 40 GB or more. Not a benchmark.NVIDIA Open Model LicenseCustom or unstated terms. The upstream license governs use.GatedGated upstream: accept the owner's terms on Hugging Face to download.
About
A 2B-parameter video world model that generates future video from text, image, or video prompts, with extra multiview checkpoints for robot and driving simulation.
Why it matters for physical AI
It lets teams generate synthetic training footage and imagine policy rollouts without collecting as much real robot video.
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.
Gated upstream. Accept the owner's terms on Hugging Face and sign in with hf auth login before downloading.
hf download nvidia/Cosmos-Predict2.5-2BUses the Hugging Face CLI (pip install -U huggingface_hub). This downloads the whole repo; add --include "<pattern>" to fetch only the files you need.
Are video world models ready for robot planning?
V-JEPA 2 plans in a learned latent space, while Cosmos generates future video that can stand in for rollouts or synthetic training data.
More robotics
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 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.