Cut AI
Realistic Film Style
5sMiniMax H3
MiniMax H3 AI Video Generator
Upload
Upload 0 image(s) · JPEG / PNG / WebP · max 10 MB each
0/7000
Required Credits20
MiniMax H3 Text, Image and Reference Modes
Write a Scene
Describe the subject, action and camera movement. Text-to-video offers six aspect ratios, from square and portrait to widescreen.
Animate Your Frames
Use one image as the first frame, or two ordered images as the first and last frames. The input frame determines the aspect ratio.
Guide with References
Use reference images, a video or audio to guide a new clip. Video-to-video generates a new interpretation from the source rather than a frame-exact edit.
Choose Duration and Detail
Generate an integer duration from 4 to 15 seconds in 768P or 2K. 768P uses 3 credits per second; 2K uses 4. A 10-second clip costs 30 or 40 credits.
How to Generate a Video with MiniMax H3
Choose a Starting Point
Select text, image, first/last frames, media references or video-to-video. Add the required media and write your direction.
Set the Output and Check the Price
Choose a duration and resolution. For text and reference modes, choose an aspect ratio. Review the displayed credit quote, then create.
Review and Save
Track the task in your workspace. Play the result, check the details, and save the clip from My Creations.
MiniMax H3 Questions
Which modes can I use?
Text-to-video, image-to-video, first/last frames, reference-to-video and video-to-video all use the same workspace. You can also select MiniMax H3 in the homepage video menu and the general video generators.
How many credits does it cost?
The rate is 3 credits per output second at 768P and 4 at 2K. Supported durations are 4–15 whole seconds, so a clip costs 12–45 credits or 16–60 credits.
Which references can I upload?
The reference workspace accepts up to nine images, one video and one audio clip. Frame mode uses exactly two ordered images. Use public media that you have permission to use.
Does image-to-video let me choose an aspect ratio?
The first frame determines the output ratio. Text and reference modes offer 21:9, 16:9, 4:3, 1:1, 3:4 and 9:16.
Can it extend or precisely edit an existing video?
Video-to-video uses the source as a reference to generate a new clip. For timeline edits or extending an existing film, use the dedicated video editor or Studio.
What happens if generation fails?
Confirmed terminal failures follow the existing credit refund flow. An uncertain provider receipt keeps the original task pending and recovers with its original request key.
Create Your First MiniMax H3 Clip
Start with a short scene, select the output and review the quote.
More AI Models
Explore another model for your next scene or image.
Seedance 2.5Longer scenes, flexible references and native audio.
Kling 4.0Guide 4–30 second scenes with image, video and audio references.
Seedance 2.0Text, image and reference-guided video creation.
Gemini OmniCreate, animate and extend reference-guided scenes.
Nano Banana 2.1Create and refine images with visual references.
Seed Audio 1.0A voice workflow guide. Speech generation is not connected.
