> ## Documentation Index
> Fetch the complete documentation index at: https://docs.anyfast.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# MiniMax-H3

> MiniMax H3 multimodal video generation and task query API.

MiniMax-H3 is MiniMax's multimodal video generation model. Through AnyFast, create a video generation task with `POST /v2/video_generation`, then query the result with `GET /v2/query/video_generation/{task_id}`.

## Key capabilities

* **Text-to-video** — Generate a video from a text prompt.
* **Image-to-video** — Use an image as the first frame, last frame, or both.
* **Multimodal reference-to-video** — Combine text with reference images, video, and audio.
* **2K output** — Request the official `2K` resolution.
* **Task-based generation** — The create endpoint returns a task ID for asynchronous status polling.

## Model ID

Use `MiniMax-H3` in the request body.

## Quickstart

Create a text-to-video task:

```bash cURL theme={null}
curl --request POST \
  --url https://www.anyfast.ai/v2/video_generation \
  --header 'Authorization: Bearer YOUR_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "MiniMax-H3",
    "content": [
      {
        "type": "text",
        "text": "史诗级太空歌剧院线预告：女舰长独自站在巨大观景窗前，最后一支舰队正在集结并跃迁离去。"
      }
    ],
    "resolution": "2K",
    "duration": 5,
    "ratio": "16:9"
  }'
```

The response contains a `task_id`:

```json theme={null}
{
  "task_id": "424010985738629"
}
```

Query the task after creation:

```bash cURL theme={null}
curl --request GET \
  --url https://www.anyfast.ai/v2/query/video_generation/TASK_ID \
  --header 'Authorization: Bearer YOUR_API_KEY'
```

## Request content

The `content` array can contain `text`, `image_url`, `video_url`, and `audio_url` parts. Use a public URL for media files; Base64 input is not supported. Set `role` when a media part has a specific purpose, such as `first_frame`, `last_frame`, `reference_image`, `reference_video`, or `reference_audio`.

## Parameter limits

* `resolution` is currently `2K`.
* `duration` accepts an integer from 4 to 15 seconds.
* `ratio` supports `adaptive`, `21:9`, `16:9`, `4:3`, `1:1`, `3:4`, and `9:16`. Text-to-video requires an explicit ratio; image-to-video uses the input image ratio.
* The total request body must not exceed 64 MB.
* Images support JPG, JPEG, PNG, WEBP, HEIC, and HEIF. Each image can be up to 30 MB, with width and height from 256 to 5760 pixels and aspect ratio from 0.4 to 2.5.
* Reference videos support MP4 and MOV with H.264 or H.265 video and AAC or MP3 audio. Each video can be up to 50 MB and 2 to 15 seconds; the total video duration is at most 15 seconds.
* Reference audio supports WAV and MP3. Each audio file can be up to 15 MB and 2 to 15 seconds; the total audio duration is at most 15 seconds.

## Query task status

The query response reports `queued`, `running`, `succeeded`, `failed`, or `expired`. When the task succeeds, use `task.content.url` to retrieve the generated video.

<Card title="API Reference" icon="code" href="/api-reference/model-api/minimax/minimax-h3">
  View the interactive create-task and query-task reference for MiniMax-H3.
</Card>

<script src="/public/feedback.js" />