MiniMaxMediumUltra

MiniMax Hailuo-03 (H3) Reference to Video

MiniMax Hailuo-03 (H3) reference-to-video via Fal.AI. Generates video from multimodal references — subject/style images, motion video clips, and audio clips, each cited in the prompt by order — keeping subjects consistent while following the referenced motion and audio.

1510 credits

per 5s 2K video — billed per second (480P 116/s, 768P 140/s, 2K 302/s, 4K 372/s); up to 5 reference images

Multimodal references: subject/style images, motion video clips, and audio clips in one request

Cite references in the prompt by order — Image 1, Video 1, Audio 1

Subject consistency across the generated take

5-15 second duration, adaptive or fixed aspect ratio (21:9 through 9:16)

480P/768P native generation, 2K/4K upscaled output

Run it right now

Test this model instantly in the Console Playground — no code required

Use with AI Assistant

Copy usage instructions for Claude, ChatGPT, or other AI

llms.txtllms.txt ↗

Quick Start

curl -X POST "https://api.core.today/v1/predictions" \
  -H "Content-Type: application/json" \
  -H "X-API-Key: cdt_your_api_key" \
  -d '{
  "model": "minimax/h3/reference-to-video",
  "input": {
    "prompt": "Image 1 is the female protagonist. Image 2 is her small dog. Keep the woman and dog consistent with their respective reference images while they walk together through a sunlit garden.",
    "reference_image_urls": [
      "https://example.com/person.jpg",
      "https://example.com/dog.jpg"
    ],
    "duration": 5,
    "resolution": "2K",
    "aspect_ratio": "adaptive"
  }
}'

Parameters

Parameter	Type	Required	Default	Description
`prompt`	string	Yes	-	Text prompt for video generation. Refer to reference assets by their modality and order: Image 1, Image 2, Video 1, Audio 1, and so on
`reference_image_urls`	array	No	-	URLs of subject/style reference images, cited in the prompt as Image 1, Image 2, ... Up to 5 images
`reference_video_urls`	array	No	-	URLs of motion/reference video clips (2-15s each, combined at most 15s), cited as Video 1, Video 2, ... Up to 3 clips
`reference_audio_urls`	array	No	-	URLs of reference audio clips (2-15s each, combined at most 15s), cited as Audio 1, ... Up to 3 clips. Cannot be the only reference input
`duration`	integer	No	5	The duration of the video in seconds (5-15). Billed per second
`resolution`	string	No	2K	The resolution of the generated video. 480P and 768P are native generation modes; 2K and 4K upscale a 768P base result 480P768P2K4K
`aspect_ratio`	string	No	adaptive	The aspect ratio of the generated video. 'adaptive' lets the model choose adaptive21:916:94:31:13:49:16
`prompt_expansion_mode`	string	No	balanced	How much effort to spend rewriting the prompt before generation. 'fast' returns in about a second; 'quality' spends up to ~30s on a richer prompt fastbalancedquality
`enable_safety_checker`	boolean	No	true	If set to true, the safety checker will be enabled
`seed`	integer	No	-	Random seed. A random seed is selected when omitted

How to Provide File Input

There are 3 ways to provide files for the reference_image_urlsreference_video_urlsreference_audio_urls parameters:

Recommended

Image URL

Pass a publicly accessible URL directly. With the Storage API (POST /v1/files/upload-url) the file uploads straight to S3 (50MB per file) and you pass the returned file_url.

{
  "model": "minimax/h3/reference-to-video",
  "input": {
    "prompt": "your prompt here",
    "reference_image_urls": ["https://example.com/image.jpg"]
  }
}

Direct Upload (Multipart)

Attach files directly to POST /v1/predictions/upload. No separate upload step, but the bytes pass through the API server so the whole request is capped at 10MB.

curl -X POST "https://api.core.today/v1/predictions/upload" \
  -H "X-API-Key: cdt_your_api_key" \
  -F "model=minimax/h3/reference-to-video" \
  -F 'input={"prompt":"your prompt here"}' \
  -F "file:reference_image_urls=@your_file.png"

See the File Upload docs for more upload methods including Presigned URLs.

Common Parameters

Common parameters used when calling POST /v1/predictions.

Parameter	Type	Required	Default	Description
`model`	string	Yes	-	Model identifier
`input`	object	Yes	-	Object containing the model-specific parameters from the table above
`output_folder`	string	No	-	Folder path for output files (max 256 chars, '..' not allowed)
`webhook_url`	string	No	-	Webhook URL to call on completion
`is_public`	boolean	No	false	If true, output files are also available via permanent public URLs

Examples

Character-Consistent Scene

Keep a person and a pet consistent with their reference images

curl -X POST "https://api.core.today/v1/predictions" \
  -H "Content-Type: application/json" \
  -H "X-API-Key: cdt_your_api_key" \
  -d '{
  "model": "minimax/h3/reference-to-video",
  "input": {
    "prompt": "Image 1 is the female protagonist. Image 2 is her small dog. Keep the woman and dog consistent with their respective reference images while they walk together through a sunlit garden.",
    "reference_image_urls": [
      "https://example.com/person.jpg",
      "https://example.com/dog.jpg"
    ],
    "duration": 5,
    "resolution": "2K",
    "aspect_ratio": "adaptive"
  }
}'

Motion Transfer

Follow the motion of a reference clip with a new subject

curl -X POST "https://api.core.today/v1/predictions" \
  -H "Content-Type: application/json" \
  -H "X-API-Key: cdt_your_api_key" \
  -d '{
  "model": "minimax/h3/reference-to-video",
  "input": {
    "prompt": "Image 1 is the dancer. Follow the motion of Video 1 while keeping the dancer consistent with Image 1.",
    "reference_image_urls": [
      "https://example.com/dancer.jpg"
    ],
    "reference_video_urls": [
      "https://example.com/dance-motion.mp4"
    ],
    "duration": 8,
    "resolution": "768P"
  }
}'