Skip to main content

MiniMax’s open-weights H3-Image generates a still from a prompt and applies up to three LoRA files, each with its own 0–4 strength, across five size tiers from 1K to 4K and 15 aspect ratios.

From $0.042/image

Workbench

Input

Text to image
154/5000
Required prose, maximum 5,000 characters; put the weight file's trigger first, then the scene. No default description or extra text fee applies. at least 1 characters · at most 5000 characters
Available ratios: 1:1, 1:2, 2:1, 1:3, 3:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 9:21, 21:9. The default generation shape is 1:1; ratio selection leaves billing unchanged.
affects price
The styled generation tier defaults to 1k, with 1.5k, 2k, 3k and 4k as higher-priced choices.
Style LoRAs1/3

Vh5tape Vhs

Accepts up to three {path, scale} entries. Point path at a direct .safetensors download, such as a Hugging Face /resolve/main/ file; scale is 0–4, default 1. A broken link returns Failed to load LoRA, with no image and no charge. The fixed LoRA increment is unchanged for zero, one or three attachments. at most 3 items
Output format
JPEGPNGWebP
The still is encoded as jpeg unless png or webp is selected; encoding does not affect the price.
Seed
Set an optional seed for a LoRA-on versus LoRA-off comparison; no fixed default or numeric limits are specified. This control is free and exact reproduction is not promised. 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
vhs-motel-red-dressExample result

Missing the model you need?

MINIMAX IMAGE LORA API

MiniMax H3-Image Open Weights LoRA API at a glance

Apply H3 video LoRAs to stills, then compare the treatment with the same seed.

LoRA capacity
3 files
Attach at most three .safetensors files using direct URLs.
Per-file scale
0–4
Each LoRA accepts scale from zero to four, with one as the default.
Tested LoRA reuse
H3 video
In our tests H3 video-model LoRAs changed the generated stills and edited photographs.
Same-seed check
On / off
Compare a request with and without its LoRA to inspect the effect.
Bad links stop the task
Fails
A missing LoRA file failed the request in our tests, yielding no image and no charge.
Five image tiers
1k–4k
The size menu includes 1.5k, 2k and 3k between its endpoints.

Results

What you can make with MiniMax H3-Image Open Weights LoRA

  • A woman in a red slip dress dances in a motel room at night, with neon light and a soft 1990s camcorder look across the frame.

    Motel dance on tape

    Neon light through a camcorder treatment

    The new 1k, 16:9 sample shows a woman in a red slip dress dancing in a motel room. With vh5tape and scale 1, the VHS LoRA gave the whole frame a soft 1990s home-video look.

    Generated motel dance sample: VHS LoRA with vh5tape at scale 1; 1k, 16:9, seed 11, PNG.

  • A couple dances in a roadside diner with a jukebox, checkered floor and grainy tape texture, with vh5tape text at lower left.

    Check the corners

    The tape look can bring its own lettering

    Our diner dance has recording grain and vh5tape lettering at lower left. That visible artifact is a reason to inspect text as well as the intended style before selecting an output.

    VHS LoRA at scale 1; 1.5k, 16:9, seed 21.

  • Two friends play a racing arcade cabinet amid neon lighting and soft VHS blur in a 1990s mall arcade.

    Another VHS setting

    A racing cabinet in a soft recording frame

    The arcade example applies the VHS treatment to two players amid neon lighting. It gives a second scene for judging how the adapter changes the appearance of people and surroundings.

    VHS LoRA at scale 1; 1.5k, 16:9, seed 11.

  • A black-and-white rubber-hose cartoon harbor shows smiling boats beneath the retained city skyline.

    Photo to cartoon

    A different drawing language for the harbor

    The Studio1939 style edit redrew the harbor in black and white with rubber-hose cartoon boats. The skyline arrangement remained recognizable in this sample.

    1930s cartoon LoRA applied to one harbor reference at 1k.

Content test

Is MiniMax H3-Image Open Weights LoRA uncensored? What we tested

Content levels tested on MiniMax H3-Image Open Weights LoRA, by mode
LevelImage editingText to image
Mild
Not tested yet
Not tested yet
Moderate
Not tested yet
Not tested yet
Strong
RendersTested Nov 11, 2026
RendersTested Nov 11, 2026
Very strong
Not tested yet
Not tested yet
Maximum
Not tested yet
Not tested yet
How we test

How each mode behaves

All modes
A broken LoRA path stops the request with Failed to load LoRA; no image is delivered and the failure is not charged.
All modes
Some LoRAs carry training-set watermark text into the image; the diner test included vh5tape in a corner.
All modes
Seed matching is a comparison aid, not a guarantee that an attached LoRA will preserve the composition.
All modes
For LoRA scenes as well, every person you describe must be an adult.
All modes
The generated downloads are kept for 14 days before deletion.

Use cases

What to build with the MiniMax H3-Image Open Weights LoRA API

  • Neon motel scene studies

    The red-dress motel example uses the VHS LoRA at scale 1 to turn a nighttime dance into a soft camcorder-style still. Use its whole-frame treatment as a reference for a tape-era mood board.

  • Restyle an existing photograph

    Our 1930s cartoon LoRA changed a photograph into rubber-hose cartoon imagery while retaining the pose. The harbor example shows a related edit that preserves the skyline arrangement.

  • Compare an adapter before scaling up

    Keep the seed and scene description steady while switching a LoRA on and off. Our VHS comparison changed the framing as well as the texture, so evaluate both before choosing a larger resolution tier.

  • Revisit a portrait setting

    Single-picture wardrobe and setting edits retained the face in our tests. Add a LoRA treatment as a separate comparison so you can judge its effect on the person and surroundings.

  • Combine source and scene references

    Two-picture composition succeeded in our tests. Put the editable source before the scene reference and name each picture in the instruction when planning a styled composite.

Prompts

MiniMax H3-Image Open Weights LoRA prompt examples you can copy

  • Text to image

    Red dress in a tape-era motel

    vh5tape, a woman in a red slip dress dancing alone in a motel room at night, laughing at the camera, neon light through the blinds, 1990s home video look.

    Why it works. An age-stated variation on the motel example connects a simple action, a specific dress and neon lighting to the tested VHS treatment. Attach the VHS file separately at scale 1.

    Try this prompt
  • Text to image

    Diner dance after closing

    vh5tape, a man and a woman in 1990s clothes dancing in a roadside diner at night, jukebox glow, checkered floor, rain on the windows, home video look.

    Why it works. The diner sample demonstrates both tape texture and an unwanted corner word. This variation makes the adults, clothing and room light clear while leaving the same artifact check relevant.

    Try this prompt
  • Text to image

    Arcade reunion

    vh5tape, two friends playing a racing arcade cabinet in a crowded 1990s mall arcade, neon glow on their faces, home video look.

    Why it works. A cabinet, wardrobe and two contrasting light sources give the VHS adapter a concrete scene to treat; the related arcade result shows the soft recording appearance.

    Try this prompt
  • Text to image

    Quiet lounge portrait

    vh5tape, a man in his forties wearing a cream linen shirt and dark trousers sits in a motel lounge at midnight, thoughtful expression, amber table lamp and distant blue neon, medium-wide home-video still.

    Why it works. This untested scene proposal holds the treatment vocabulary steady while changing the subject and mood. Compare it with the adapter removed before adjusting its scale.

    Try this prompt
  • Image editing

    Harbor in an old cartoon

    Studio1939 style, <Picture 1> redrawn as a 1930s black-and-white rubber-hose cartoon harbor, boats bobbing with smiling faces, keep the skyline layout.

    Why it works. The harbor edit is a measured example of a photo changing drawing style while keeping its skyline arrangement. The source token identifies what receives the treatment.

    Try this prompt
  • Image editing

    Portrait with a scene reference

    vh5tape, place the woman in her forties from <Picture 1> into the motel lounge shown in <Picture 2>, wearing a navy silk robe over a cream dress, seated beside the amber lamp with a calm smile; keep her face recognizable and apply a 1990s recording look.

    Why it works. This proposed edit assigns separate roles to the portrait and room. Use photos you have the right to use, adults only, with no impersonation; compare the styled output for identity and framing changes.

    Try this prompt

Pricing

MiniMax H3-Image Open Weights LoRA API pricing

One image starts at $0.042 at 1K and $0.222 at 4K.

Get started

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

  1. Step 1

    Choose generation or editing

    Start from a written scene for generation, or supply the picture to revise before any supporting references. Editing accepts up to nine input pictures.

  2. Step 2

    Attach a downloadable LoRA

    Point path at the .safetensors file itself, using a Hugging Face /resolve/main/ link when appropriate. Start with scale 1; each file accepts 0–4 and the request can contain up to three files.

  3. Step 3

    Describe the scene and treatment

    Put the selected LoRA trigger ahead of a detailed English description of the wardrobe, action, light and surroundings. For an edit, map the source to <Picture 1> and assign later references their own numbers.

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

FAQ

MiniMax H3-Image Open Weights LoRA API FAQ

About MiniMax H3-Image Open Weights LoRA

A LoRA file supplies a visual treatment, but its effect can reach beyond surface texture. Use the motel dance, diner and harbor samples to decide what you want to change, then compare your own scene with the attachment enabled and removed. Keep the seed, canvas and description steady while evaluating one adjustment. Review the edges for carried-over lettering and check whether the composition still fits your brief. Choose a larger resolution tier after that review, or use the Open Weights sibling when the scene needs no adapter.

All MiniMax models21

Related models and next tasks