20% offBlack Friday11d 03:20:21 left

How to write MiniMax H3 Image prompts

H3 Image does best with a plain-English description of the whole picture: what is in it, where, in what light, shot how. This MiniMax H3 Image prompt guide shows how to build one, how it differs from a MiniMax H3 video prompt, when to turn Add detail on or off, and how to keep text in the image exact, with real prompts and the images they made on this site.

Add photos to edit them (up to 16)

New accounts get 20 free credits · See pricing

Not affiliated with MiniMax · Your images are never used to train AI · Privacy policy

H3 Image prompts vs MiniMax H3 video prompts

Most MiniMax H3 prompt guides online are written for the video model. H3 Image is a separate still-image model, and much of a video prompt does nothing here. The difference in one table:

In the promptH3 ImageMiniMax H3 video
What you describeOne frame: what is in it, where, in what lightShots that change over time
CameraA fixed position and a lens lookHow the camera moves within and between shots
Words in quotesPrinted in the picture as textUsually lines of dialogue, spoken aloud
ReferencesUp to 16 images, called <Picture 1>, <Picture 2>…Images, and in some workflows video or voice

Reusing a video prompt? Keep the description of the first frame, drop the camera moves, dialogue and sound effects, and write the camera as a fixed position.

Five parts of a good prompt

Take this film still. Its prompt is one long sentence, but each phrase does a different job.

A 35mm film photograph of a colossal derelict starship half-buried in desert sand, a lone farmer leading a mule past it in the foreground for scale, hazy late-afternoon light, real-world materials and weathering, rust streaks and sand drifts against the hull, muted film colours, fine grain
Made on this site with MiniMax H3 Image · 2K · 4:3 · Add detail on. Every phrase quoted on the right is copied from this prompt.
Prompt

A 35mm film photograph of a colossal derelict starship half-buried in desert sand, a lone farmer leading a mule past it in the foreground for scale, hazy late-afternoon light, real-world materials and weathering, rust streaks and sand drifts against the hull, muted film colours, fine grain

1
Subject
Name the main thing and its size or state, not only its type.“a colossal derelict starship half-buried in desert sand”
2
Scale and foreground
Put a person or a familiar object in frame, so a big thing reads as big.“a lone farmer leading a mule past it in the foreground for scale”
3
Light
Describe it the way a photographer would: time of day, weather, how hazy or hard it is.“hazy late-afternoon light”
4
Materials and wear
Say what the surfaces are made of and how they have aged.“real-world materials and weathering, rust streaks and sand drifts against the hull”
5
Kind of picture and look
Open with what kind of picture it is, and close with the colour and the grain.“A 35mm film photograph … muted film colours, fine grain”

Add detail: on or off?

MiniMax’s model card says H3 Image expects a detailed English description of the whole image, and that its API can turn a short prompt in any language into one. On this site that step is the Add detail switch. Source: Hugging Face model card

3D game environment concept art: an ancient stone amphitheatre half-swallowed by a forest, giant tree roots breaking through the seating tiers, collapsed arches and fallen columns, moss and ivy, shafts of morning light through the canopy, realistic game-engine render, no characters1K · Add detail on

On: when you describe a mood or a scene

You write the idea and the model fills in light, materials and detail. This game concept art was made that way from a single sentence.

Prompt

3D game environment concept art: an ancient stone amphitheatre half-swallowed by a forest, giant tree roots breaking through the seating tiers, collapsed arches and fallen columns, moss and ivy, shafts of morning light through the canopy, realistic game-engine render, no characters

A café chalkboard menu hanging on a brick wall, hand-lettered in white and yellow chalk. Title at the top: "THE CORNER CUP". Two columns. Left column under "COFFEE": "Espresso 3.00", "Flat White 4.20", "Oat Latte 4.80", "Cold Brew 4.50". Right column under "BAKERY": "Butter Croissant 3.50", "Banana Bread 3.80", "Cardamom Bun 4.00". At the bottom: "Open daily 7am – 4pm · Free refills on filter coffee". Small chalk doodles of a coffee cup and a leaf. Warm morning light, photo.2K · Add detail off

Off: when the words must be exact

Your prompt is used exactly as written, and a run is 10–20 seconds faster. This menu was made that way, so the quoted words were not rewritten.

Prompt

A café chalkboard menu hanging on a brick wall, hand-lettered in white and yellow chalk. Title at the top: "THE CORNER CUP". Two columns. Left column under "COFFEE": "Espresso 3.00", "Flat White 4.20", "Oat Latte 4.80", "Cold Brew 4.50". Right column under "BAKERY": "Butter Croissant 3.50", "Banana Bread 3.80", "Cardamom Bun 4.00". At the bottom: "Open daily 7am – 4pm · Free refills on filter coffee". Small chalk doodles of a coffee cup and a leaf. Warm morning light, photo.

Getting exact text in the image

Text is where prompts fail first. These six rules come from repeated runs on this site, and each note says what went wrong without it.

  1. 1

    Put every word in double quotes and say where it goes

    Write the title, each line and the punctuation exactly as it should appear, and say where it sits. Our chalkboard menu, with a title, two columns and a footer line all spelled out this way, came out right at 2K on the first run.

  2. 2

    Keep the word count down

    A recipe card with about ten lines came back with four misspelt words and a repeated step. Cut to six lines, and asked for neat printed letters with every word spelled as written, both versions were correct.

  3. 3

    Spell out the layout

    An infographic that included a small table gained a garbled row and a broken table. Asking for five equal columns, each with one icon and one caption, and no other text, gave two clean versions.

  4. 4

    End with “No other text.”

    Without it, a poster added a line of its own: a year we never asked for.

  5. 5

    Say what each label points to

    For labelled diagrams, name the part for every label, such as “Woofer” on a large round paper-cone driver. Asking only for “a label on each part” gave five garbled labels out of six.

  6. 6

    Turn Add detail off

    Add detail rewrites your prompt first, and the rewrite can change quoted words or turn Chinese characters into English.

A clean flat infographic poster on a light beige background. Title at the top: "Pour-Over Coffee in 5 Steps". Below it, five equal columns from left to right, each with one simple line icon and one caption: "1. Heat water to 94°C", "2. Grind 20 g coffee", "3. Rinse the filter", "4. Bloom 30 seconds", "5. Pour to 320 g". No other text.2K · Add detail off
Rule 3 in practice: five equal columns, one icon and one caption each, no other text. 2K · 4:3 · Add detail off.
Prompt

A clean flat infographic poster on a light beige background. Title at the top: "Pour-Over Coffee in 5 Steps". Below it, five equal columns from left to right, each with one simple line icon and one caption: "1. Heat water to 94°C", "2. Grind 20 g coffee", "3. Rinse the filter", "4. Bloom 30 seconds", "5. Pour to 320 g". No other text.

Camera, light and scale

Four habits that held up in our tests

  1. 1

    Say where the camera is, in plain words. “Worm’s-eye view, camera rests on the sidewalk, looking up” worked; “two feet above the ground” was ignored.

  2. 2

    Describe light like a photographer: “hazy late-afternoon light”, “overcast morning”, “shafts of sun through water”.

  3. 3

    Give a huge thing a person for scale. The farmer and the mule are what make the starship above read as colossal.

  4. 4

    For a film look, ask for “a 35mm film photograph” with real-world materials and weathering, so it doesn’t come out looking like a game render.

Every image on this page was generated on this site with MiniMax H3 Image, with no retouching. The settings are tagged on each image, and the prompt shown with it is the exact text used.

Before you press Generate

  1. 1

    Does the prompt say what kind of picture it is: a photo, a render, a poster, a diagram?

  2. 2

    Are the light and the camera described in plain words?

  3. 3

    Is every word that must appear in the image inside double quotes?

  4. 4

    Is Add detail off if exact text matters, and on if you only have an idea?

Where this comes from

  • The model card describes H3 Image as expecting a detailed English description of the whole image, with prompt expansion for short prompts.

    Hugging Face model card
  • MiniMax’s own announcement of H3 Image, for what the model is and what it is built to do.

    MiniMax announcement

Guides

More H3 Image prompts

  • MiniMax H3 Image example: Game art

    Game art

    Describe a level, a ruin or a prop set and get engine-style concept art that sticks to what you wrote.

    Game asset prompts
  • MiniMax H3 Image example: Character sheets

    Character sheets

    Upload one photo or a few screenshots of a character, or describe one in a sentence, and these character sheet prompts give you a model sheet with labelled turnaround views and close-ups.

  • MiniMax H3 Image example: Text in images

    Text in images

    Menus, letters, labels, signs and infographics with the exact words you typed.

    Text-in-image prompts
  • MiniMax H3 Image example: Portraits

    Portraits

    Close-ups with real skin, stray hair and natural light, and candid frames that look like someone just picked up a camera.

    Portrait prompts
  • MiniMax H3 Image example: Product shots

    Product shots

    Start from one photo of your product and get it in a new scene, from a new angle, as a technical breakdown or in a finished ad, with its shape, colour and details kept the same.

    Product photo prompts
  • MiniMax H3 Image example: Film stills

    Film stills

    Frames that look shot on film rather than rendered: real materials, natural light and a camera you can place.

    Cinematic prompts
  • MiniMax H3 Image example: Many references

    Many references

    Upload up to 16 images at once (people, props, a place, a painting style, a colour card) and say in one prompt how they fit together.

  • MiniMax H3 Image example: Chinese calligraphy

    Chinese calligraphy

    Write the prompt in Chinese, quote a poem or an idiom, and H3 Image paints the scene and writes the characters into it: vertical calligraphy, red seals, spring couplets on a door.

All H3 Image prompts

Make your first H3 image

Your text goes into the prompt box at the top of the page, with the credit cost shown before you run it.