MiniMax H3
Complex character locomotion, cloth physics, and natural facial gestures.
Key Features of MiniMax H3
Engineered for precision, speed, and creative flexibility.
Three Modes
Text, frame, and multi-reference generation in one model.
2K Output
Above-FHD resolution for detailed renders.
Reference Sets
Three or more reference images drive subject consistency.
Technical Specifications
MiniMax H3 is a triple-mode video model: text-to-video, first/last-frame image-to-video, and a reference mode that animates three or more images.
Per-second billing: 5 credits at 768p, 8 at 2K.
How to use MiniMax H3
- 01
Provide Text or an Image
Describe the scene you want to animate, or upload a still frame to use as the starting point of your clip.
- 02
Set Duration & Quality
Choose clip length, resolution, aspect ratio, and generation mode for your video.
- 03
Generate & Export
Render the clip and preview it instantly. Download the video for social platforms, campaigns, or storyboards.
MiniMax H3 FAQ
Explore more models and tools
Ready to start creating?
Claim 20 free credits upon sign up. No credit card required.
Launch AI Workspace