MiniMax H3, sound and references in one pass.
MiniMax H3 is the model to reach for when the shot has to match assets you already own. It takes up to 9 reference images and up to 3 reference video clips — the only engine on ScriptMotion that reads more than one clip at a time. Output lands at 2K with dialogue and ambient audio baked into the same generation.
Generated on ScriptMotion using MiniMax H3— no editing.
Nine images and three clips of context
Feed H3 up to 9 reference images in one brief, plus up to 3 reference video clips. Every other engine that reads reference video takes a single clip — H3 is the only one that can follow a whole set of existing edits at once.
Dialogue and ambience, not just music
H3 writes dialogue and ambient sound into the same generation pass, delivered as a standard H.264 video with an AAC audio track — no separate voice or sound step.
2K output, 21:9 through 9:16
Render at 768P for drafts or 2K for delivery, across five aspect ratios — 16:9, 9:16, 1:1, 3:4, and 21:9 — including cinematic 21:9 for banners and 3:4 for portrait feeds.
How to create with MiniMax H3
Pick the Custom layout
MiniMax H3 is selected directly in the Custom layout, where you also set duration anywhere from 4 to 15 seconds and choose your aspect ratio.
Load your references
Attach up to 9 reference images for characters, products, and scenes, and add up to 3 reference video clips when you want H3 to follow existing motion or pacing.
Generate with sound at 2K
H3 returns a clip with dialogue and ambient audio already mixed in. Credit cost scales with resolution and duration and is shown before you generate.
MiniMax H3 specs
Every aspect ratio you need
Generate MiniMax H3videos in the formats each platform expects — sized automatically when you pick your layout.
MiniMax H3 FAQ
Generate with MiniMax H3 — and 14 more models.
Every model on every paid plan, full-speed credits, commercial use included. From $39/mo.