Skip to main content

MiniMax H3-Image, the hosted release of MiniMax’s image model, turns a detailed English prompt into a PNG, JPEG or WebP still at 1k, 2k or 4k, with 15 aspect ratios and short, readable lettering.

From $0.036/image

Workbench

Input

Text to image
184/5000
Required scene description of at most 5,000 characters; no preset text or additional prompt charge. at least 1 characters · at most 5000 characters
The default is 1:1; alternatives are 15 ratios from 3:1 to 1:3, among them 1:1, 16:9, 9:16, 21:9 and 4:5, or custom for an exact pixel size set with width and height. Shape selection does not alter the generation rate.
affects price
Pixel tier: 1k by default, about 1 MP; 2k about 4 MP; 4k about 16 MP. The 1k and 2k rates match, while 4k costs three times as much. A custom size is billed at this tier: up to 4.5 MP at 1k or 2k, larger at 4k.
Prompt expansion
true by default: MiniMax first rewrites your scene into a longer English description. Set false to send the prompt word for word, for instance when a quoted label must stay exactly as typed. No price effect.
Output format
JPEGPNGWebP
Choose png, jpeg or webp for the generated file; the default is png and format leaves pricing unchanged.
Seed
Optional whole number from 0 to 2,147,483,647 with no price effect. MiniMax only reuses the starting point while prompt expansion is off, so pair a fixed seed with enable_prompt_expansion false for side-by-side tests; matching pictures are not guaranteed. no lower than 0 · no higher than 2147483647
Safety checkerOff

SpicyAPI does not run a safety checker on this model. Your prompt and media go to the model as written, and the model itself decides what comes back.

Checking your account before enabling submission and pricing.

Output

Ready for a new run
room-1207-noteExample result

Missing the model you need?

MINIMAX HOSTED IMAGE API

MiniMax H3-Image API at a glance

Three output sizes, short readable labels, and edits guided by up to sixteen pictures.

One rate for two sizes
1k / 2k
The 2k tier has the same generation charge as 1k.
Reference capacity
16
Editing accepts between one and sixteen uploaded pictures.
Larger canvas
4k
In our tests a square 4k result measured 4096 x 4096.
Measured 1k processing
12–15 s
In our tests model time at 1k was about 12–15 seconds, excluding queueing and transfer.
Labels and signage
Short
In our tests MIDNIGHT JAZZ and NOIR 07 were readable in the returned images.

Results

What you can make with MiniMax H3-Image

  • A woman in a black silk slip lies across white hotel sheets in morning light, holding a handwritten note reading "Room 1207, midnight".

    Handwritten message

    Room 1207, written into the scene

    The 2k, 3:2 result shows a woman in a black silk slip across hotel sheets, holding a note that legibly reads "Room 1207, midnight". The requested words survived inside a photographic scene with soft morning light. Use this sample to inspect a brief handwritten line; it does not establish reliable rendering of long passages.

    2k, 3:2 generated sample: the handwritten note legibly carries the requested words "Room 1207, midnight".

  • A singer in a blue sequined gown stands at a vintage microphone, with brass players and a glowing MIDNIGHT JAZZ sign behind her.

    Short text in context

    Lettering that belongs to the club

    Our club image contains a readable MIDNIGHT JAZZ sign, and the separate perfume test rendered NOIR 07 on its label. These are short phrases placed within a scene. When supplying your own copy, quote the phrase and identify the surface where it belongs, then check the returned letters individually.

    2k club scene at 16:9, measuring 2720 x 1536.

  • A golden retriever rests on a wool rug beside a mustard sofa, round walnut table and green plant in a sunlit room.

    Reference roles

    Give the room and subject separate jobs

    The furnished-room example makes the purpose of two references easy to follow: one supplies the space, the other the dog. Our composition tests successfully moved an object or person from one picture into another scene. Name the sources by upload order and explain which elements of the setting should remain.

    1k interior edit using two reference pictures.

  • A terracotta canvas sneaker with cream trim and a gum sole rests on a concrete floor.

    A focused revision

    Describe the change around a source object

    The sneaker example pairs a requested colorway with instructions to keep the source shape and floor. Follow that change-and-retain structure when drafting a product edit. Review the object against its source after generation; an instruction to preserve a detail is a goal to inspect, rather than a guarantee of an unchanged image.

    One shoe reference with a revised colorway.

More clips made with MiniMax H3-Image

  • A woman in a black satin slip dress leans on the shoulder of a smiling man in an open white shirt at a candlelit hotel bar.
    Two-portrait edit: image 1 and image 2 combined into one bar scene.
  • A man wearing a navy turtleneck sits beside a rain-streaked window in soft daylight.
    4k portrait at 2:3; returned dimensions 3360 x 5024.
  • A chef in a white jacket cooks noodles in a flaming wok at a rainy night-market stall lit by strings of lights.
    One source picture relocated to a night market.
  • An amber perfume bottle with a NOIR 07 label stands on black marble.
    1k square product photograph with requested lettering.

Content test

Is MiniMax H3-Image uncensored? What we tested

Content levels tested on MiniMax H3-Image, by mode
LevelText to imageImage editing
Mild
RendersTested Nov 11, 2026
Not tested yet
Moderate
RendersTested Nov 11, 2026
RendersTested Nov 11, 2026
Strong
RendersTested Nov 11, 2026
RendersTested Nov 11, 2026
Very strong
SoftenedTested Nov 11, 2026
RendersTested Nov 11, 2026
Maximum
RendersTested Nov 11, 2026
Not tested yet
How we test

How each mode behaves

All modes
Long passages of text are not established by our short-sign and label tests.
All modes
Negative prompt, guidance and sampling steps are not exposed; a seed repeats a starting point only with prompt expansion switched off.
All modes
When composing a request, every person you describe must be an adult.
All modes
Generated files remain available for 14 days and are deleted afterward.

Use cases

What to build with the MiniMax H3-Image API

  • Portraits built around a written note

    Describe an adult subject, a quiet room and the exact words on a card. The Room 1207 example returned its requested handwritten message legibly at 2k and 3:2, giving a concrete starting point for a scene where the wording matters alongside the pose and light.

  • Short labels in product imagery

    Keep the required copy brief and tie it to a named surface, such as a bottle label. Our NOIR 07 result had readable lettering; the MIDNIGHT JAZZ scene also showed a clear sign. Those observations support checking short text rather than planning a page of prose inside the picture.

  • Wardrobe and location revisions

    Bring a permitted adult portrait to the editor and describe a new outfit or setting, with a separate instruction to retain the person. Our single-image changes kept the face recognizable. Inspect each returned edit before treating it as an accepted continuation of the source.

  • A subject placed in another scene

    Use one input for a person or object and another for the destination environment. Our two-picture compositions combined those roles successfully. Numbering the uploads lets an instruction identify the subject to move and the scene that should receive it.

  • Wide scenes and square compositions

    Choose the output shape for the intended placement, then set the pixel tier separately. Our measurements include 2720 x 1536 at 2k and 16:9, 4096 x 4096 at 4k and 1:1, and 5472 x 3072 at 4k and 16:9; the tier does not prescribe one fixed width.

Prompts

MiniMax H3-Image prompt examples you can copy

  • Text to image

    A message in the morning light

    A woman in a black silk slip lying across rumpled white hotel sheets in soft morning light, holding up a handwritten note that reads "Room 1207, midnight", a sly smile, photorealistic.

    Why it works. This variation on the hotel example names an adult subject, the message and the light, with the words attached to a visible card.

    Try this prompt
  • Text to image

    Portrait beside a rainy window

    Editorial portrait of a man in his forties with a salt-and-pepper beard and a navy wool turtleneck, sitting by a large window on a grey morning, soft natural light, medium format film, natural skin texture.

    Why it works. Age, clothing, mood and window light form a connected portrait brief, inspired by the existing rainy-window sample.

    Try this prompt
  • Text to image

    An amber bottle with a short label

    A minimalist perfume bottle on a black marble plinth, the glass label reads "NOIR 07" in thin gold serif letters, single soft key light, luxury product photography.

    Why it works. The requested phrase is brief and assigned to a specific surface, following the label test rather than asking for extended copy.

    Try this prompt
  • Text to image

    Evening conversation on a terrace

    An adult couple in their forties sits close at a small terrace table after sunset. She wears a deep green evening dress and he wears an open-collar linen shirt. They share a quiet smile over coffee, with warm lantern light, a blue evening sky and distant city windows behind them.

    Why it works. A defined relationship, wardrobe and lighting direction give a new scene a coherent mood without relying on extra controls.

    Try this prompt
  • Image editing

    A chef in a different setting

    Move the chef in image 1 to a night-market street food stall in the rain, cooking noodles in a wok under string lights. Keep his face, beard and white jacket.

    Why it works. The source role, new setting and preservation goals are stated separately for the existing chef-edit scenario.

    Try this prompt
  • Image editing

    A dog in the furnished interior

    Furnish the empty room in image 1 with a mustard velvet sofa, a round walnut coffee table and a large green plant, and put the dog from image 2 lying on a wool rug in front of the sofa. Keep the window light.

    Why it works. Two numbered references have distinct jobs, making the requested combination easy to audit against the source pictures.

    Try this prompt
  • Image editing

    A warmer shoe colorway

    Repaint the sneaker in image 1 in a terracotta and cream colorway with a gum rubber sole, keep the shape and the concrete floor.

    Why it works. This suggested revision places the intended color changes beside a concrete list of details to retain.

    Try this prompt

Pricing

MiniMax H3-Image API pricing

One image starts at $0.036 at 1K and $0.108 at 4K.

Get started

How to use the MiniMax H3-Image API in three steps

  1. Step 1

    Choose generation or editing

    Begin with text-to-image for a scene described in words, or image-to-image when existing pictures should guide the result. For people in editing inputs, use photos you have the right to use, depict adults and avoid impersonation.

  2. Step 2

    Describe the intended picture

    Write connected English prose covering the subject, wardrobe, mood, setting and light within 5,000 characters. In an edit, name image 1, image 2 and later uploads, then specify the changes and the details to retain.

  3. Step 3

    Set shape, size and file type

    Select a canvas from fifteen ratios and choose 1k, 2k or 4k; 1k and 2k share a generation rate, while 4k costs three times as much. PNG is the default, with JPEG and WebP available at the same format-independent rate.

Parameter reference and code samples are in the API and Readme tabs at the top of this page.

FAQ

MiniMax H3-Image API FAQ

About MiniMax H3-Image

Plan an H3-Image request around the part of the picture you need to control: a short written message, an adult portrait in a new setting, or a subject placed into a reference scene. The Room 1207 sample makes the lettering easy to inspect, while the editing examples show how numbered sources divide a composition into clear roles. Select the ratio deliberately and compare 1k with 2k using their shared generation rate before choosing 4k. Review text and source details in the returned image, then keep the accepted result with its prompt and settings for the next revision.

All MiniMax models21

Related models and next tasks