Now on Mixio · Multimodal video

MiniMax H3,
directed inside Mixio

Hailuo 3.0 — text, images, video, and audio read as one brief. Up to 2K, clips to 15 seconds, with dialogue and sound generated in the same pass — routed through the platform that keeps your characters and continuity locked.

Up to 2K
Native resolution
5–15s
Clip length
24 FPS
Film cadence
Stereo
Native audio

Capabilities

One model that reads the whole brief.

H3 takes text, images, video, and audio as a single context and returns a finished audio-visual clip. Mixio wraps a production layer around it, so one generation becomes part of a consistent whole.

Omni reference

Feed up to 9 images, 3 clips, and 3 audio files in one run — a shot can inherit a face, a motion, and a voice at once.

Native stereo audio

Dialogue, effects, and room tone are generated with the picture, in the same pass — no silent clip to score later.

Instruction-based editing

Name the one element to change and the rest of the frame holds, so a take you already like survives the revision.

Voice cloning & transfer

A reference recording hands a character a new voice, and a supplied line can replace what was said on camera.

Character consistency

Face, styling, motion, and voice carry across shots when the same reference set drives each generation.

Multi-shot in one clip

Several shots sit inside a single 5–15s generation — title sequences and shot-reverse-shot land in one run.

Prompt → Output

One brief. A clip that arrives with sound.

MiniMax H3 · 2K

Prompt

A perfume bottle catches the last light in a rooftop glasshouse. Slow push-in, 16mm grain, one sustained cello note as the sun clears the frame.

Three ways in

Text, two frames, or a full reference set.

Text-to-video

Type the brief, get the shot

Write subject, action, camera, light, and sound as one brief and pick from six aspect ratios — 21:9 through 9:16. A finished clip returns with audio already in it.

First & last frame

Two frames, filled in

Hand H3 a start frame and an end frame and it composes the motion between them, following the aspect ratio of the images you upload.

Omni reference

Give every reference a job

Attach up to 12 files — images, clips, audio — and assign each a role: this face, that location, this motion, that voice. H3 reads them as one brief.

Showcase

Vertical, with the sound already in it.

Real H3 renders in 9:16 — dialogue, effects, and room tone generated alongside every frame.

Vertical drama
Product
Lifestyle
Documentary
Music

Use cases

What teams ship with H3.

Brand films & commercials

TV spots and premium brand films where a reference set fixes the talent, the product, and the closing mark — and the film texture is specified, not left to chance.

Vertical drama & dialogue

Short-form 9:16 drama built on close coverage and shot-reverse-shot. The line is spoken in the same pass as the picture, so performance and delivery arrive together.

Product & e-commerce

Product films built from a still of the real object — a full reveal into macro detail and a held wide. The reference holds the shape while camera and light do the selling.

Motion design & titles

Title work where type is its own layer: credits type on, diagrams assemble, and the music cue is timed to the beat you name rather than added later.

Game & interface concepts

Menus, HUDs, and interaction demos timed beat by beat across the clip, so panels slide and a world loads in the order you wrote, not one the model guessed.

Stylized & animated work

Paper-cut, stop-motion, and stylized character films. An identity reference holds costume, proportion, and finish from one cut to the next.

Developers

One key. Every model. H3 included.

Call H3 — with its omni references and native audio — from the same endpoint you use for Kling, Veo, Runway, and Seedance. Switch models by changing one field. No new SDK, no separate billing.

  • Single API key routes to every supported model
  • Attach image, video, and audio references in one request
  • MCP workflows for Claude, Codex, Cursor, and Windsurf
  • Outputs land in your production library and timeline
curl https://api.mixio.pro/v1/generations \
  -H "Authorization: Bearer $MIXIO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "minimax-h3",
    "mode": "omni-reference",
    "prompt": "0-5s she walks the benches, 5-11s she lifts the bottle into the light",
    "references": [
      { "role": "location", "image_url": "https://cdn.mixio.pro/glasshouse.png" },
      { "role": "talent",   "image_url": "https://cdn.mixio.pro/lead.png" },
      { "role": "voice",    "audio_url": "https://cdn.mixio.pro/vo.wav" }
    ],
    "aspect_ratio": "16:9",
    "resolution": "2k",
    "duration": 12
  }'

MiniMax H3 on Mixio

Everything you need to know about generating with Hailuo 3.0.

Put H3 to work in your production.

Free to start · One key routes every model