Skip to content
HiAPI Docs
HiAPI
Playground Get API Key
English
English 中文

MiniMax H3 API

POST /v1/tasks

All models are called through the Unified Async API POST /v1/tasks endpoint; only the input fields differ (see input parameters below).

Model summary

Model name minimax-h3
Type Video generation (text-to-video)
Endpoint POST /v1/tasks
Pricing See HiAPI Pricing

MiniMax H3 video generation API: one model ID for text-to-video, frame control, and multimodal references, with native 2K output from 4 to 15 seconds.

Production guidance

Production guidance
  • For production, pass callback.url at the top level of the request body so HiAPI can notify your service when the task reaches a terminal state.
  • GET /v1/tasks/:id is better for local debugging, low-volume jobs, or fallback reconciliation if a callback is missed.
  • Use callback.when=final. Both success and fail are terminal states, so your service should deduplicate by taskId.

Best suited for

Native 2K text-to-video

Generate 4-15 second 2K clips from a prompt for ads, social content, and concept previews.

promptdurationaspect_ratio
First and last frame control

Provide a first frame, or both first and last frames, to control how the shot starts and resolves.

first_frame_imagelast_frame_image
Multimodal references

Combine reference images, videos, and audio to guide subject, motion, camera, or rhythm.

image_urlsvideo_urlsaudio_urls

Request parameters

model string required

Fixed value minimax-h3.

example minimax-h3
input object required

Business parameters. Put MiniMax H3-specific configuration here.

prompt string required

Video content prompt. Required for every generation mode, with a maximum length of 7000 characters.

duration integer optional

Generated video duration in whole seconds.

default 5
resolution enum optional

Output resolution. MiniMax H3 currently generates at native 2K only.

default 2K enum: 2K
aspect_ratio enum optional

Output aspect ratio. Text-to-video defaults to 16:9; frame-guided video follows the input image; reference mode may use adaptive.

default 16:9 enum: 21:916:94:31:13:49:16adaptive
watermark boolean optional

Whether to add an AIGC watermark to the generated video.

default false
first_frame_image string optional

Public image URL used as the first frame. Do not combine frame controls with reference media fields.

last_frame_image string optional

Public image URL used as the last frame. Requires first_frame_image.

image_urls string[] optional

Reference image URLs for multimodal reference-to-video. HiAPI accepts up to five images. Do not combine with first/last-frame controls.

video_urls string[] optional

Reference video URLs for multimodal reference-to-video. Up to three clips, each 2-15 seconds and no more than 15 seconds total.

audio_urls string[] optional

Reference audio URLs. Up to three clips and no more than 15 seconds total. Audio cannot be used alone; include at least one reference image or video.

callback object optional

Optional callback configuration. When set, HiAPI notifies your service when the task reaches a terminal state.

url string required

Required when callback is set; HTTPS URL that receives terminal task notifications.

example https://your-domain.com/hiapi/callback
when enum optional

Callback trigger timing. Use final.

default final enum: final

Example requests

Minimal text-to-video

Pass only prompt. Defaults are 5 seconds, 2K, 16:9, and no watermark.

Request body
{
  "model": "minimax-h3",
  "input": {
    "prompt": "A paper kite drifting gently above a green field under soft daylight, stable camera, no text."
  }
}
First and last frame video

Control the shot boundaries. last_frame_image must be used with first_frame_image.

Request body
{
  "model": "minimax-h3",
  "input": {
    "prompt": "The camera slowly pushes forward as morning light transitions naturally into golden sunset.",
    "duration": 6,
    "first_frame_image": "https://example.com/first-frame.jpg",
    "last_frame_image": "https://example.com/last-frame.jpg"
  }
}
Multimodal reference video

Use up to five reference images and optionally add video or audio references. Do not combine these with frame controls.

Request body
{
  "model": "minimax-h3",
  "input": {
    "prompt": "Keep the reference subject consistent and follow the camera rhythm of the reference clip for a polished product showcase.",
    "duration": 8,
    "aspect_ratio": "16:9",
    "image_urls": [
      "https://example.com/product.jpg"
    ],
    "video_urls": [
      "https://example.com/camera-motion.mp4"
    ],
    "audio_urls": [
      "https://example.com/rhythm.mp3"
    ]
  }
}

Getting the result

  1. The response returns a taskId immediately without waiting for generation to finish.
  2. In production, prefer waiting for callback.url to receive the terminal notification. For local debugging, poll GET /v1/tasks/:id.
  3. When status=success, download the generated video from output[].url.
  4. When status=fail, fix the request based on the returned error instead of retrying the same invalid payload.

FAQ

Which generation modes does MiniMax H3 support?

The same minimax-h3 model ID supports text-to-video, first/last-frame video, and multimodal reference-to-video. The mode is selected by the frame or reference fields present in input.

How is it billed?

HiAPI bills by output video second. MiniMax H3 currently exposes native 2K output only. Check the pricing page for the current rate before submitting.

What are the reference media limits?

Up to 5 images; up to 3 videos, each 2-15 seconds and 15 seconds total; and up to 3 audio clips with 15 seconds total. Audio cannot be used alone and requires at least one reference image or video.

Can frame controls and reference media be combined?

No. Frame controls and image_urls/video_urls/audio_urls select different modes and must be sent in separate requests. A last frame also requires a first frame.

Next steps