# MiniMax H3 (https://docs.fuser.studio/docs/nodes/video/minimax-h3)

_Provided by [Fal](https://fal.ai) — [Learn More](https://fal.ai/models/minimax/h3/text-to-video)._

## Inputs [#inputs]

- **Prompt** (`textarea`) — The text prompt to generate a video from. In reference mode, cite provided media as Image 1, Video 1, Audio 1, and so on.

- **Start Image** (`image`) — The image used as the first frame of the video (image-to-video mode).

   Optional: `true`

- **End Image** (`image`) — The image used as the last frame of the video. Only available when a start image is provided.

   Optional: `true`

- **Reference Images** (`image`) _(accepts multiple)_ — Subject and style reference images, cited in the prompt as Image 1, Image 2, and so on. Reference images, videos, and audio clips must add up to at most 12 files.

   Optional: `true`

- **Reference Videos** (`video`) _(accepts multiple)_ — Motion reference video clips (2–15 seconds each, combined duration at most 15 seconds), cited in the prompt as Video 1, Video 2, and so on.

   Optional: `true`

- **Reference Audio** (`audio`) _(accepts multiple)_ — Reference audio clips (2–15 seconds each, combined duration at most 15 seconds), cited in the prompt as Audio 1, Audio 2, and so on. Audio cannot be the only reference — provide at least one reference image or video with it.

   Optional: `true`

- **Duration** (`slider`) — The duration of the generated video in seconds.

   Optional: `true` • Default: `5` • Minimum: `5` • Maximum: `15`

- **Resolution** (`dropdown`) — Video resolution. 768p for faster generation, 2K for balance, 4K for maximum quality.

   Optional: `true` • Default: `2K` • Accepts: `768p`, `2K`, `4K`

- **Aspect Ratio** (`dropdown`) — The aspect ratio of the generated video. "Auto" adapts to the references in reference mode and defaults to 16:9 for text-to-video. Image-to-video always follows the start image.

   Optional: `true` • Default: `16:9` • Accepts: `Auto`, `21:9`, `16:9`, `4:3`, `1:1`, `3:4`, `9:16`

- **Prompt Expansion** (`toggle`) — Let the model expand and enrich the prompt automatically before generation.

   Optional: `true` • Default: `true`

## Outputs [#outputs]

- **Generated Video** (`video`) — The generated video from the model.