Motion study
“A dancer turns slowly through shafts of window light in an empty rehearsal studio, dust drifting, the camera orbiting at waist height.”
video model
Generate fast AI video with MiniMax H3 Max
MiniMax H3 Max is a speed-focused, post-trained variant of the H3 video model, generating 5 to 15 second clips at 480P or 768P with strong prompt adherence — often in a few seconds.
$0.08 / second at provider cost for members — see the full price breakdown
What you need to start
Start with a text prompt, or add a start image. An optional end frame can guide the final composition.
Loading...
Original results
MiniMax H3 Max examples made in Made
These examples use media produced for Made. Prompts are included when the production record allows them.
Image-to-video
“Bring the scene to life with a slow orbital camera move, natural fabric motion, and drifting dust in the light.”
Creative scene
“Close-up of espresso pouring into a glass cup in warm morning light, steam curling, a slow push-in on the crema.”
Start-frame control
“Use this image as the opening composition.”
Scene direction
“Establish the framing and light before describing the motion.”
Composition brief
“Plan the shot as a still, then hand the motion to the model.”
Where MiniMax H3 Max fits best
Best uses
- Very fast motion and camera-direction iteration
- Animating a still image with an optional end frame
- Short social clips in wide, square, or portrait formats
- Low-cost drafts before a premium video model pass
Limits to know
- Output tops out at 768P, so final masters that need 1080p should be regenerated on a higher-resolution model or upscaled.
- Prompt expansion is on by default and can reinterpret terse prompts; set precise motion language for exact control.
Specifications in Made
These are the controls exposed by Made, which may be narrower than a provider’s full API.
- Aspect ratios
- 21:9, 16:9, 4:3, 1:1, 3:4, 9:16
- Resolution
- 480P, 768P
- Duration
- 5–15 seconds
- Output formats
- MP4
- Reference images
- 2
- Processing
- Typically generates a 5-second clip in a few seconds; longer clips take proportionally longer.
How to use MiniMax H3 Max in Made
- 1
Choose text or image input
Write a prompt on its own, or add a start image — and optionally an end frame — to anchor the clip.
- 2
Describe motion precisely
Name the subject movement, camera path, and pacing; precise language matters with prompt expansion enabled.
- 3
Set format and length
Pick an aspect ratio, 480P or 768P, and a duration between 5 and 15 seconds.
- 4
Iterate quickly
Generation is fast enough to explore several motion directions before committing to a final take.
Compare relevant models
video
Seedance 2.5
Seedance 2.5 is ByteDance’s newest video model in Made, generating clips up to 30 seconds in a single pass with native audio at 480p to 1080p — the longest single-shot ceiling in the Made library.
Explore model →video
WAN 3.0 Prime
WAN 3.0 Prime is the premium tier of Alibaba’s WAN 3.0 video family, generating clips from 2 to 30 seconds with synchronized audio at up to 1080p, including 30 fps image-to-video.
Explore model →video
WAN 2.2 Turbo
WAN 2.2 Turbo is Made’s fast iteration option for short text-to-video and image-to-video clips, with portrait and landscape output controls.
Explore model →Continue with a related tool
AI Video Generator
Turn any image into a captivating video. Upload a photo, describe the motion you want, and watch AI bring it to life with realistic movement, cinematic camera work, and professional quality.
Open tool →AI Image Generator
Create scroll-stopping product photos and marketing visuals in seconds. Describe your shot, pick your style, and get professional images ready for your store, social feed, or ad campaign.
Open tool →It is a post-trained variant of the MiniMax H3 video model tuned for prompt adherence and aesthetics, optimized to generate short clips in seconds.