Generation modes
text to videoimage to video
MiniMax · video
MiniMax H3 AI Generator
MiniMax H3 is an omni-modal generation model built for high-impact commercial video, complex instructions and native audio-visual creation. The current integration supports 2 generation modes, 480p, 768p and 3–15 second clips.
Resolutions
480p768p
Durations
3–15s
Aspect ratios
16:99:161:14:33:421:99:21
Model selection guide
How to get the best from MiniMax H3
MiniMax H3 is an omni-modal generation model built for high-impact commercial video, complex instructions and native audio-visual creation.
MiniMax H3 strengths
- Strong instruction following for commercial concepts
- Native audio-video workflow in the official model family
- Good fit for motion, text and brand-sensitive scenes
Best use cases for MiniMax H3
- Product advertising and branded content
- Dialogue, ambience and sound-led concepts
- Motion-reference and multi-shot experiments
MiniMax H3 prompt tips
- Write dialogue and sound cues beside the relevant shot
- Specify brand text exactly and keep it short
- Separate camera, subject motion and environmental motion
Editorial note: capabilities are summarized from official releases, public examples and the parameters currently integrated by VidVid. The generator controls are the source of truth for available options.
Official capability overview
What makes MiniMax H3 different
H3 is a general-purpose omni-modal video model. Its distinction is not resolution alone: reference understanding, picture and sound generation share one workflow.
Native audio-video generation
Generates 32 kHz stereo ambience and dialogue with the picture, reducing post-production and alignment work.
Omni-modal references
One API can understand text, up to nine images, three video clips and three audio clips for stronger subject and brand consistency.
4–15 seconds, up to 2K
The official API supports every whole-second duration from 4 to 15 at 24 FPS, with 768P and 2K output.
Complex instructions and text
A strong fit for multi-shot narrative, motion transfer, product advertising and shots that need readable brand text.
Specifications: official MiniMax H3 release and API documentation, checked 2026-08-27.
Credit examples
Credits update with selected parameters and are shown before generation.
3s · 480p
24 credits
4s · 480p
32 credits
5s · 480p
40 credits
Create with MiniMax H3
Choose a generation mode
Add a prompt or reference image
Set parameters and generate
Frequently asked questions
What generation modes does MiniMax H3 support?
text to video, image to video
What is MiniMax H3 best for?
Product advertising and branded content, Dialogue, ambience and sound-led concepts, Motion-reference and multi-shot experiments
Which durations and resolutions does MiniMax H3 support?
3–15s; 480p, 768p
Can I see the exact cost before generating?
Yes. Credits update with duration and resolution and appear on the Generate button.