MiniMax H3
MiniMax H3 : generate a video that keeps the subject from a reference image, driven by a text prompt. Supports 2K, 5-15s.
MiniMax H3 Video
The MiniMax H3 video generation model produces high-quality videos from multimodal content (text, images, video, audio). It offers an out-of-the-box REST inference API with 2K direct output, first/last-frame control, and multimodal reference-based video generation.
Base URL
https://api.icreat.aiAuthentication
All API requests must be authenticated with an API Key. You can obtain an API Key from the console.
export ICREAT_API_KEY="your-api-key-here"HTTP Request Headers
import os
API_KEY = os.environ.get("ICREAT_API_KEY")
headers = {
"Content-Type": "application/json",
"Authorization": "Bearer " + API_KEY,
}Protect your API Key
Never expose your API Key in client-side code or public repositories. Use environment variables or a backend proxy.
Code Examples
Image and video generation uses a two-step async flow: submit a task to get task_id, then poll via query task result; the response includes status and result ([] while processing; on SUCCEEDED, result holds resources and costUSD is present). The examples below use the same task_id across both steps.
1. Submit Task
Send a generation request to the submit endpoint. On success, the response body returns task_id — save it for polling and result retrieval.
2. Query Task Result (Poll)
Use the task_id from submit to poll progress (repeat until terminal). The response includes status and result: result is [] while processing; on SUCCEEDED, result holds resources and costUSD is included; FAILED means the task failed.
Input Schema
Submit Task — Input
The following parameters are accepted in the submit request body.
Total: 6 Required: 3 Optional: 3
Multimodal input array. Each element is differentiated by type (text / image_url / video_url / audio_url); use role to annotate purpose. Each request must include a non-empty text item.
Video resolution.
Generated video duration in seconds.
Min: 4 Max: 15
Aspect ratio of the generated video. Recommended for text-to-video mode.
Callback URL for task status changes.
Whether to add an AIGC identification watermark to the output video.
Query Task Result — Input
Total: 1 Required: 1 Optional: 0
The task ID returned from the submit endpoint.
Output Schema
Submit Task — Output
Total: 1
Unique identifier for the async task. Use this ID to poll status and retrieve results.
Query Task Result — Output
Total: variable
Current task status. result is usually [] until success; on SUCCEEDED, result holds resources and costUSD is present.
Generated resources. Empty array while processing or on failure; array of objects on success.
Task cost in USD. Present only when status is SUCCEEDED.
LLM Prompt
The Markdown below is an LLM-friendly prompt you can paste into AI assistants (e.g. Cursor, ChatGPT) to help them understand this model's API, call flow, and key parameters. Use Copy for AI or copy from the code block below.
# minimax/h3-video
> The MiniMax H3 video generation model produces high-quality videos from multimodal content (text, images, video, audio).
## Overview
Use the iCreat two-step async task API: submit a generation request, then poll the query task result endpoint; on success read resources from `result` (includes `costUSD`).
## API Info
- **Base URL**：`https://api.icreat.ai`
- **Submit endpoint (POST)**：`/v1/task/submit/minimax/h3-video`
- **Query result endpoint (POST)**：`/v1/task/result`
- **Model ID**：`minimax/h3-video`
- **Auth**：`Authorization: Bearer ${ICREAT_API_KEY}`
## Call Flow
1. **Submit**: POST submit path with body per Input Notes; response `{ "task_id": "..." }`
2. **Query result**: POST `/v1/task/result` with `{ "task_id": "..." }`; response includes `status` and `result` (`[]` while processing); on `SUCCEEDED`, `result` holds resources and `costUSD` is present; read `url` or `download_url` when `type` is `Video`
### Input Notes
- Request body centers on a multimodal `content` array plus resolution/duration params
- `role` (optional): Purpose annotation, e.g. `reference_image` for image-to-video.
- `resolution` (required): Video resolution.
- `duration` (required): Generated video duration in seconds.
- `ratio` (optional): Aspect ratio of the generated video. Recommended for text-to-video mode. (default `adaptive`)
- `callback_url` (optional): Callback URL for task status changes.
- `aigc_watermark` (optional): Whether to add an AIGC identification watermark to the output video. (default `false`)
### Output Notes
- Poll: read `status`; `result` is `[]` while processing
- Success: `result` is `[{ "type": "Video", "url": "...", "download_url": "..." }]` plus `costUSD`
## Notes
- Use the same `task_id` across both steps; `result` is `[]` while processing — keep polling
- `FAILED` is terminal — check request parameters or reference media