API reference · NVIDIA

nvidia/cosmos-3-super/image-to-video

Integrate this model through SandBase's unified API, with production-ready schemas and examples.

VIDEOAsyncOpen model

Production endpoint

Send your first request

OpenAI-compatible endpoint with unified authentication and usage tracking.

POSThttps://api.sandbase.ai/v1/run

Model IDnvidia/cosmos-3-super/image-to-video

Input Schema

12 parameters · 2 required · 10 optional

Parameter	Type	Required	Description
`image`	string	Required	URL of the conditioning first-frame image for the video.
`prompt`	string	Required	Text prompt describing the motion and scene of the video to generate. · Min length: 1 · Max length: 4096
`seed`	integer	Optional	The same seed and prompt given to the same model version will produce the same video every time.
`num_frames`	integer	Optional	Number of frames to generate. More frames yield a longer video. · Min: 5 · Max: 189 · Default: 189
`aspect_ratio`	string	Optional	The aspect ratio of the generated image. · Options: 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16 21:916:93:24:35:41:14:53:42:39:16
`guidance_scale`	number	Optional	Classifier-free guidance scale. Higher values increase prompt adherence at the cost of diversity. · Min: 0 · Max: 20 · Default: 6
`frames_per_second`	integer	Optional	Frames per second of the output video. · Min: 4 · Max: 60 · Default: 24
`agentic_early_stop`	boolean	Optional	Stop the agentic loop early when the critic score clears the strict quality threshold. · Default: true
`num_inference_steps`	integer	Optional	Number of denoising steps. More steps yield higher quality but take longer. · Min: 1 · Max: 50 · Default: 28
`agentic_max_iterations`	integer	Optional	Maximum agentic prompt stages when agentic generation is enabled. · Min: 1 · Max: 3 · Default: 2
`enable_agentic_generation`	boolean	Optional	Enable the iterative Cosmos agentic loop: prompt upsampling, candidate video generation, VLM critique of sampled frames, and prompt rewrite. Each candidate is a full render, so this is substantially slower and costlier than a single generation. · Default: false
`agentic_samples_per_iteration`	integer	Optional	Candidate videos to generate and judge per agentic iteration. The best candidate advances to the next rewrite stage. · Min: 1 · Max: 3 · Default: 2

Output Schema

Field	Type	Description
`id`	string	Unique identifier for the generation task
`status`	string	Task status: pending, running, completed, failed, timeout
`model`	string	Model used for the generation
`outputs`	array	Array of output items
`outputs[].url`	string	URL of the generated artifact
`outputs[].content_type`	string	MIME type (e.g. image/png, video/mp4)
`error`	object \| null	Error details if failed, null on success
`error.type`	string	Machine-readable error type code
`error.message`	string	Human-readable error description

Async Workflow

This model uses asynchronous execution. Submit a request and poll for the result.

Submit — POST to /v1/run, receive an id
Poll — GET /v1/run/{id} until status is completed, failed, or timeout
Retrieve — Read outputs from the completed response

Code Examples

Ready-to-run snippets

# Step 1: Submit
curl -X POST https://api.sandbase.ai/v1/run \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
  "model": "nvidia/cosmos-3-super/image-to-video",
  "image": "https://storage.googleapis.com/falserverless/example_inputs/hunyuan_i2v.jpg",
  "prompt": "The camera slowly pushes in as the subject turns their head toward the light, hair drifting in a gentle breeze, dust motes floating through warm afternoon sun.",
  "num_frames": 189,
  "guidance_scale": 6,
  "frames_per_second": 24,
  "agentic_early_stop": true,
  "num_inference_steps": 28,
  "agentic_max_iterations": 2,
  "enable_agentic_generation": false,
  "agentic_samples_per_iteration": 2
}'

# Step 2: Poll result (replace <id>)
curl https://api.sandbase.ai/v1/run/<id> \
  -H "Authorization: Bearer YOUR_API_KEY"

← Back to Cosmos 3 Super Image to Video