Skip to main content
NVIDIA
Explore
Models
Skills
Blueprints
GPUs
Docs
Help Center
Getting Started
  1. Create and verify your account to unlock full access to NVIDIA NIM APIs.
ResourcesDeveloper ForumsContact Support
FAQs
  • Terms of Use
    Privacy Policy
    Your Privacy Choices
    Contact

    Copyright © 2026 NVIDIA Corporation

    NVIDIA

    cosmos3-nano

    DownloadableFree Endpoint

    Generates physics-aware videos from text prompts or an image prompt for physical AI development.

    • Physical AI
    • autonomous vehicles
    • image-to-world
    • robotics
    • text-to-world
    • Synthetic Data Generation
    Download and Post-Train

    Cosmos3 Nano

    Cosmos | Code

    Model Overview

    Cosmos3 Nano is an NVIDIA world foundation model for generating physics-aware videos from text prompts or an image prompt. The API exposes the Cosmos3 generator path and returns an MP4 video encoded as a base64 string.

    The hosted Preview API deployment includes safety guardrails and SynthID watermarking for generated video outputs.

    Model Developer: NVIDIA

    Input/Output Specifications

    Input

    • Text prompt.
    • Optional image input as base64, data URI, or public URL.
    • Optional generation controls including seed, resolution, frame count, fps, steps, guidance scale, and negative prompt.

    Output

    • Video.
    • MP4 container.
    • Base64 encoded response field: b64_video.

    Usage

    Send requests to the model endpoint with a JSON body:

    {
      "prompt": "A robot moves through a clean industrial warehouse.",
      "resolution": "720_16_9",
      "num_output_frames": 189,
      "seed": 42
    }
    

    The response contains:

    {
      "b64_video": "<base64 encoded mp4>"
    }
    

    License

    This model is released under the NVIDIA Open Model License. For a custom license, contact cosmos-license@nvidia.com.

    Under the NVIDIA Open Model License, NVIDIA confirms:

    • Models are commercially usable.
    • You are free to create and distribute Derivative Models.
    • NVIDIA does not claim ownership to any outputs generated using the Models or Derivative Models.

    Ethical Considerations

    NVIDIA believes Trustworthy AI is a shared responsibility and has established policies and practices to enable development for a wide range of AI applications. Developers should work with their internal teams to ensure the model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

    For more detailed information, see the Explainability, Bias, Safety and Security, and Privacy cards.

    On this page

    1. Model Overview
    2. Input/Output Specifications
    3. Usage
    4. License
    5. Ethical Considerations