Video Generation

MiniMax H3

MiniMax's latest H3 (Hailuo-03) video model. Text-to-video, image-to-video (first/last frame), and reference-to-video (reference images, videos, audio) with cinematic 2K output.

Typical generation time: 1-6 minutes

Model

Text describing the video (required, up to 7000 characters). Use Shot 1: / Shot 2: for multi-shot sequences.

First Frame Image

Maximum file size: 10 MB

Starting frame for image-to-video. Cannot be combined with reference media.

Last Frame Image

Maximum file size: 10 MB

Ending frame (requires First Frame). First & last frame image-to-video.

Reference Images

Maximum file size: 10 MB

Up to 9 reference images (reference-to-video). Cannot be combined with first/last frame.

Reference Videos

Maximum file size: 100 MB

Up to 3 reference videos (2-15s each) for motion/style transfer.

Reference Audios

Maximum file size: 50 MB

Up to 3 reference audios (2-15s each) for voice/timbre.

Video length in seconds (4-15). Cost scales with duration.

Output resolution. 2K costs more.

Required for text-to-video (cannot be adaptive). Ignored for image-to-video (uses the image ratio).

Explore example prompts and their results.

  • Text-to-Video (2K)Video

Similar Models

Community