NVIDIA Cosmos

World Foundation Models for Physical AI

One model family that sees, reasons, simulates, and acts.

Try it in your browser · Quickstart · Find your path · Model Family

   ![Physics-aware generation: Newton's cradle](assets/demos/physics_newton_cradle.gif) 
  [Physics-aware generation](cookbooks/cosmos3/generator/audiovisual/run_with_vllm_omni.ipynb)  

A Newton's cradle in motion: momentum transfer rendered with physical fidelity

   ![World-scenario transfer: control layout to photoreal driving video](assets/demos/transfer_worldscenario.gif) 
  [Transfer](cookbooks/cosmos3/generator/transfer/run_video_transfer_with_vllm_omni.ipynb)  

World-scenario control layout → photoreal driving video

   ![Robot policy executing a manipulation task](assets/demos/policy_screwdriver.gif) 
  [Action policy](cookbooks/cosmos3/generator/action/run_policy_with_vllm_omni.ipynb)  

Policy run: "put the screwdriver and the glove in the purple container"

   ![Driving simulation for autonomous vehicles](assets/demos/driving_sim_falling_rocks.gif) 
  [Simulation](cookbooks/cosmos3/generator/audiovisual/run_with_vllm_omni.ipynb)  

Generate synthetic data and create simulation for autonomous driving

   ![Forward dynamics egocentric rollout](assets/demos/fd_egocentric_repair_poses.gif) 
  [Action-conditioned World Model](cookbooks/cosmos3/generator/action/run_fd_with_vllm_omni.ipynb)  

Forward dynamics: egocentric rollout from input camera + hand pose

   ![World Reasoner: dashcam hazard anticipation](assets/demos/reasoner_driving_hazard.gif) 
  [World reasoning](cookbooks/cosmos3/reasoner/run_with_vllm.ipynb)  

Reason in complex real-world scenarios: a rolling ball means a child or pet may follow

What's new

• [Jul 2026] Cosmos3-Edge released — the 4B tier for on-device, real-time deployment (Jetson AGX Orin / Thor / RTX Pro 6000).
• [May 2026] Cosmos 3 released: HF collection · Technical Report.

What is Cosmos?

NVIDIA Cosmos is an open platform for building physical AI applications — robots, autonomous vehicles, and smart infrastructure — providing better data, better environment, better starting point, and better tooling for physical AI developers. Cosmos 3, the current model family, is a suite of omnimodal world models built on a unified Mixture-of-Transformers architecture (technical report).

One model, two surfaces:

This repository is the home of the models: everything for exploring, running, and evaluating Cosmos. For model training (SFT, LoRA, RL, distillation etc.), go to Cosmos Framework. Use Cosmos Curator for data curation, and Cosmos Evaluator for model output evaluation.

Find your path

I want to…

Go to

Time

See it work — zero install

Video generation, visual reasoning

1 min

Generate my first video

Quickstart ↓

10 min

Reason over images & video

Reasoner notebook

10 min

Serve an OpenAI-compatible API

Serving setup guide — vLLM, vLLM-Omni, or NIM

30 min

Post-train on my own data — SFT, distillation, RL

Cosmos Framework, then evaluate here

hours

Explore runnable notebooks

Cookbooks

browse

Evaluate a model

Evaluation suites — PAIBench, Physics-IQ, VLMEvalKit

hours

Check latency & throughput

Benchmarks

browse

Generate your first video

Before running the code, request access to nvidia/Cosmos-1.0-Guardrail and accept its access conditions. Once access is granted, log in below with a Hugging Face read token from the same account. Logging in alone does not grant access.

uv venv --python 3.13 --seed --managed-python && source .venv/bin/activate
uv pip install --torch-backend=auto \
  "diffusers @ git+https://github.com/huggingface/diffusers.git" \
  accelerate av cosmos_guardrail huggingface_hub imageio imageio-ffmpeg \
  torch torchvision transformers
uvx hf@latest auth login   # Authenticate for the gated Guardrail repository

import torch
from diffusers import Cosmos3OmniPipeline
from diffusers.utils import export_to_video

pipe = Cosmos3OmniPipeline.from_pretrained(
    "nvidia/Cosmos3-Nano", torch_dtype=torch.bfloat16, device_map="cuda"
)
video = pipe(prompt="A mobile robot navigates a warehouse aisle and stops at a shelf.").video
export_to_video(video, "first_video.mp4", fps=24)

First run downloads the 16B checkpoint; diffusion steps are compute-heavy, so long step times are normal. Full options, image/sound modes, and every other backend: audiovisual cookbooks · setup issues: environment setup guide.

Models

Cosmos 3 ships as three base models — every deployment tier, one omnimodal architecture:

Example checkpoints in the cosmos3-examples collection are post-trained variants of the base models. They demonstrate what post-training with Cosmos Framework can specialize Cosmos for — they're capability demos, not part of the product line:

Full I/O specs, generation settings, and supported action embodiments: model reference.

Repository map

cosmos/
├── cookbooks/          # runnable notebooks for every capability (start here to explore)
│   └── cosmos3/        # generator (audiovisual · action · transfer) · reasoner + prompt guide
├── evaluation/         # quality benchmark suites: PAIBench, Physics-IQ, RBench, UniGenBench, VLMEvalKit
├── docs/
│   └── reference/      # lookup: model reference
├── assets/             # brand + demo media
└── README.md           # you are here

Training, optimization, and deployment tooling lives in Cosmos Framework.

Platform

Cosmos 3 runs on Diffusers, Transformers, vLLM, vLLM-Omni, SGLang, TensorRT-LLM, and NIM — pick a backend in the environment setup guide.

Community & contributing

Questions and ideas → Discussions. Bugs → Issues. Code → CONTRIBUTING.md.

Limitations & safety

Cosmos 3 can produce artifacts in long, high-resolution, or physically complex outputs (temporal inconsistency, object morphing, implausible dynamics). Safety-critical applications need additional validation and system-level safety analysis. Generation ships with guardrails on by default.

Citation & license

@techreport{nvidia2026cosmos3,
  title  = {Cosmos 3: Omnimodal World Models for physical AI},
  author = {{NVIDIA Cosmos Team}},
  year   = {2026},
  url    = {https://research.nvidia.com/labs/cosmos-lab/cosmos3/technical-report.pdf}
}

Source code and models are released under OpenMDW-1.1. Custom licensing: . This project may download third-party open source software; review those licenses before use.