MiniMax H3 Reference to Video
A familiar subject. A new scene with movement and sound.



Whoops! Total... Intentional drift! You guys love that, don't you? Ashen? Hey, it's me. Keep this line open, yeah?
How it works
Try it now
01Add your reference
Drop it in, or pick one of yours.
- 02
Say what you want
A sentence or two is enough.
03Make it
Start creating, then keep working with HowWorks.
Explore the possibilities

MiniMax H3 at a glance
- Good at
- Best at changing one thing and keeping the rest. Native audio. 5–15s, up to 4K.
- Modes
- Text to Video, Image to Video, First & Last Frame, Reference to Video
- Length
- 5–15 s (presets 5, 10, 15 s)
- Resolution
- 480p, 720p, 2K, 4K
- Aspect ratios
- 16:9, 9:16, 1:1, 4:3, 3:4, 21:9
- Image-led framing
- May follow the input image; check the output frame
- Sound
- Always on
- Reference images
- Up to 9
About
A product photograph or character image can guide a new scene without defining its opening composition. MiniMax H3 reference to video supports up to nine image references and includes generated audio. Give each image a clear purpose: the object shape, location or visual treatment. Write one main action and the sounds that accompany it, then choose 5–15 seconds and output through 4K. The model uses a single Standard tier, leaving you to focus on input roles, framing and scene direction. Inspect how the result used the references before choosing it.
Questions
What can reference images guide?
Use them to show a subject, object, setting or visual treatment. Explain which details matter and describe the new shot. Choose Image to Video when one supplied image must become the opening frame.
How many references can MiniMax H3 use?
The model supports up to nine image references. A small set with clearly described roles is a useful starting point; every added image should contribute information the scene needs.
Does this workflow include audio?
Yes. MiniMax H3 always includes audio. Describe desired ambience or action sounds and inspect the soundtrack for unwanted speech or noise.
Which duration, output and tier apply?
Choose 5–15 seconds and 480p, 720p, 2K or 4K output. MiniMax H3 has one Standard tier; the Turbo option belongs to MiniMax H3 Max.
Will the subject be identical to the reference?
Inspect recognizable features, proportions, materials and lettering throughout the clip. Input support describes accepted guidance, not a guarantee of exact preservation.










