Cosmos 3 Super Text2Image 4Step is NVIDIA's image generation model. A distilled variant of NVIDIA's Cosmos text-to-image model that generates images in only four diffusion steps for faster inference.
Specifications
Canonical IDnvidia-cosmos-3-super-text2image-4step
TypeImage Generation
StatusActive
CreatorNVIDIANVIDIA
Input ModalitiesText
Output ModalitiesImage

Capabilities

Input1/5
Text
Image·
Audio·
Video·
PDF·
Output1/5
Text·
Image
Audio·
Video·
Embedding·
Capabilities0/13
Reasoning·
Adaptive Reasoning·
Function Calling·
Parallel Function Calling·
Structured Outputs·
Native JSON Schema·
Web Search·
URL Context·
Computer Use·
Code Execution·
File Search·
Prompt Caching·
Assistant Prefill·

Versions

VersionReleasedContextInput / 1MOutput / 1MStatus
Cosmos 3 Super Text2Image 4StepCurrent
Cosmos 3 SuperAvailable
Cosmos 3 Super Image2VideoAvailable
Cosmos 3 Super ReasonerAvailable
Cosmos 3 Super Text2ImageAvailable
Cosmos 3 Super Text2Image AgenticAvailable
Cosmos3 Super Image2Video 4StepAvailable

Other Models

ModelTierReleasedContextInput / 1MOutput / 1M
Cosmos 3 NanoNano
Cosmos
Cosmos Reason1 7B

Model IDs

cosmos3-super-text2image-4step
nvidia-cosmos-3-super-text2image-4step