Video generation
Available now

MiniMax H3

by MiniMax

MiniMax's multimodal video model.
2K, native stereo audio, clips up to 15s.

MiniMax H3

Key features

A canoe glides through a flooded forest at first lightTry now

Technical specifications

5 to 15s

One generation, with multiple shots possible inside it.

Up to 2K

1440-pixel short edge; a 768p mode is announced as coming.

24 FPS

The cadence film and broadcast are shot at.

3 modes

Text-to-video, first and last frame, and omni reference.

Use cases

Brand films and commercials

Trailers, TV spots, and premium brand films, one of the production categories MiniMax builds the model's examples around.

Vertical drama

Short-form drama in 9:16, where dialogue, performance, and shot-reverse-shot coverage carry a 15 second hook.

Product and e-commerce

Product showcases, feature demos, and performance-marketing assets built from a still of the real product.

Motion design and titles

Title sequences, typography, and VFX packages, with the type animation directed as its own layer of the brief.

Game and interface concepts

Game UI, web UI, interaction demos, and feature walkthroughs, timed beat by beat across the length of the clip.

Stylized animation

Claymation, anime promos, and 3D character films, with an identity reference holding the design across shots.

Prompt examples

Brand film

A bottle catches the last light in a rooftop glasshouse

Edit prompt

Vertical drama

Two sisters share a silence in a hospital corridor at night

Edit prompt

Title sequence

Orbital diagrams ink themselves in beside monospace credits

Edit prompt

Product film

A tonearm settles into the groove of a spinning record

Edit prompt

Game interface

Inventory panels open on a dock as the archipelago loads

Edit prompt

Paper-cut animation

A kite pulls a child across layered paper hillsides

Edit prompt

Simple pricing

Get started for free today, with the option to upgrade or cancel anytime.

Basic

$9/ month
billed as $0 per year

900 monthly credits

1 user only

All models

Workflows

Standard

$24/ month
billed as $0 per year

3200 monthly credits

1 user only

All models

Workflows

Pro

$45/ month
billed as $0 per year

6200 shared monthly credits

1 user

+ up to 4 more at extra cost

All models

Workflows

Pro Max

$170/ month
billed as $0 per year

24000 shared monthly credits

1 user

+ up to 9 more at extra cost

All models

Workflows

Enterprise

For higher limits

Custom

pricing and billing terms

High-volume credits
Custom seat limits
All models
Workflows
Pricing Gradient

Free

For playing around

$0

forever free

Up to 20 credits
1 user only
Limited models
Workflows
Compare plan details

FAQs

What is MiniMax H3?
MiniMax H3 is MiniMax's general-purpose multimodal video model. It reads text, images, video, and audio together as one context, then returns a 5 to 15 second clip at up to 2K and 24fps with native stereo audio. It runs three ways: from a text prompt alone, from a first and last frame, or from a mixed set of references.
Is MiniMax H3 the same as Hailuo 3.0?
Yes. MiniMax H3 is the official model name, and Hailuo 3.0 is the name it is widely referred to by, after Hailuo AI, the MiniMax app it runs in. You will also see it written Hailuo 03 and Hailuo-03. All of these point at one model, which follows Hailuo 2.3 in the same line. Note that it is not related to Kling O3, a different model from Kuaishou that is easy to confuse with the Hailuo 03 spelling.
Is MiniMax H3 open weight?
MiniMax describes H3 as an open-weight, general-purpose multimodal video model, and has said the weights will be released. Until they are, the model is used through the Hailuo AI app, the MiniMax Hub desktop app, the MiniMax Open Platform API, and hosting providers that have picked it up.
How long can a MiniMax H3 video be?
Between 5 and 15 seconds per generation. Multiple shots can sit inside that single clip, so a title sequence or a drama beat with shot-reverse-shot coverage is produced in one run rather than cut together from separate generations.
What resolution and frame rate does MiniMax H3 output?
1440p, described as 2K, at a fixed 24 frames per second. The 1440 figure is the short edge, so a 16:9 through 9:16 clip measures 1440 on its narrow side, and wider formats land near 3.7 megapixels, roughly 2976 by 1248 at 21:9. A 768p mode is announced as coming, and its output can be upscaled to 1440p.
What aspect ratios does MiniMax H3 support?
Six: 21:9, 16:9, 4:3, 1:1, 3:4, and 9:16. Text-to-video and omni reference runs let you pick, and omni reference adds an auto mode where H3 chooses the framing from the references you supply. A first and last frame run instead follows the aspect ratio of the image you upload.
What can you feed MiniMax H3 as a reference?
Up to 9 images, 3 video clips, and 3 audio files, capped at 12 files in total for one generation. Reference video and reference audio each run 2 to 15 seconds per clip and 15 seconds in total. Audio cannot be sent on its own; it has to travel with at least one image or video.
How do I use MiniMax H3 on Morphic?
Select MiniMax H3 in the prompt bar, switch to video, and write the shot as a short brief covering the subject, the action across the clip, the camera, the light, and the sound you want. Attach any reference images, clips, or audio the shot needs to hold, then generate. Takes land on the Canvas, so you can compare H3 against another model before picking one.
How is MiniMax H3 different from Hailuo 2.3?
Hailuo 2.3 tops out at 1080p and roughly 10 seconds from a prompt or a single image. MiniMax H3 moves to 2K, stretches to 15 seconds, and adds the parts that change how you work: mixed image, video, and audio references, native stereo audio, voice transfer, and edits you request in a sentence instead of re-rolling the shot.

More about MiniMax H3

MiniMax H3: complete guide, references, editing, and audio

The complete MiniMax H3 guide: omni references, native audio, instruction-based editing, and the brief structure that gets the most out of the model.

MiniMax H3 prompt library: video briefs that work

A growing library of MiniMax H3 video prompts with example clips. Each is a full shot brief with reference roles, timed beats, and directed sound.

Other models

Seedance 2.5

ByteDance

ByteDance's next-generation video model. Up to 30s native clips, 50 references, native audio.

Kling 4.0

Kling

Kuaishou's next Kling flagship, expected soon. Longer clips and sharper detail are on the way.

Flux 3

Black Forest Labs

Black Forest Labs' world model for image, video, and audio, with sound that matches the action.

Seedream 5.0 Pro

ByteDance

Grounded editing and multi-layer separation. Dense infographics and native text in 14 languages.