MiniMax · Image, text and reference to video
MiniMax H3 — native 2K across three input modes
Image, text or reference-driven video at 2K, 4 to 15 seconds, 21:9 through 9:16.
Try MiniMax H3What is MiniMax H3?
MiniMax H3 covers three input modes under one model: image-to-video with first, last or both keyframes; text-to-video with aspect ratios from 21:9 to 9:16; and reference-to-video taking up to 9 reference images, 3 reference videos and 3 reference audio clips. All three output native 2K at 4 to 15 seconds. In image-to-video the frame shape adapts to the input image rather than being set separately.
Available on Floniks
Credit cost per render, read live from the model catalogue. All models draw from one balance — no per-model subscription.
- MiniMax H3 Image to Video75 credits
- MiniMax H3 Text to Video75 credits
- MiniMax H3 Reference to Video75 credits
What MiniMax H3 is best for
- Work that needs native 2K without an upscaling pass
- Vertical and ultrawide output from the same model
- Reference-driven video with many source images
How to use MiniMax H3
Choose the input mode
Image, text or reference. Reference mode accepts up to 9 images, 3 videos and 3 audio clips.
Set duration and aspect
4 to 15 seconds. Text-to-video takes an explicit aspect ratio; image-to-video adapts to your input image.
Render at 2K
Output is native 2K, so no separate upscale step is needed for most delivery targets.
Frequently asked questions
- Does MiniMax H3 output 2K natively?
- Yes, across all three input modes — no separate upscaling pass required.
- What aspect ratios does MiniMax H3 support?
- Text-to-video covers 21:9 through 9:16. In image-to-video the aspect adapts to the input image.
- How many references does the reference-to-video mode take?
- Up to 9 images, 3 videos and 3 audio clips.
