Introduction
The hosted generator accepts a prompt plus optional shape, pixel tier and file-format choices; the default result is a 1k PNG.
| Field | Value |
|---|---|
| Input | Text |
| Output | Image |
| Publisher and model | MiniMax; H3-Image, for still-image generation and reference editing |
| Required prompt | Detailed English description; up to 5,000 characters |
| Resolution tiers | 1k (default, about 1 MP), 2k (about 4 MP), 4k (about 16 MP) |
| Generation rate rule | 1k and 2k share a rate; 4k costs three times that rate |
| Canvas ratios | 15 ratios from 3:1 to 1:3, among them 1:1, 16:9, 9:16, 21:9 and 4:5, or a custom width and height in pixels |
| Output encoding | PNG by default; JPEG and WebP also supported; format does not change pricing |
| Editing inputs | 1–16 images; reference them as image 1, image 2 and onward in upload order |
| Accepted input formats | JPEG, PNG, WebP, HEIC and HEIF |
| Input dimensions | 128–8192 pixels on each edge; long edge at most four times the short edge |
| Shape when omitted | Text-to-image: 1:1; editing: chosen by the model from the references and prompt |
| Controls not exposed | Negative prompt, guidance and sampling steps |
| Measured pixel dimensions | 1k 1:1: 1024 x 1024; 2k 16:9: 2720 x 1536; 4k 1:1: 4096 x 4096; 4k 16:9: 5472 x 3072 |
Key features & highlights
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.
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.
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.
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.
Request parameters
| Parameter | Required | Description |
|---|---|---|
prompt | Yes | Required scene description of at most 5,000 characters; no preset text or additional prompt charge. at least 1 characters · at most 5000 characters |
aspect_ratio | No | 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. |
width | No | Exact output width in pixels, 468–7,745, used only when |
height | No | Exact output height in pixels, 468–7,745, required alongside width whenever |
resolution | No | 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. |
enable_prompt_expansion | No | true by default: MiniMax first rewrites your scene into a longer English description. Default: |
output_format | No | Choose png, jpeg or webp for the generated file; the default is png and format leaves pricing unchanged. |
seed | No | Optional whole number from 0 to 2,147,483,647 with no price effect. |
Pricing
Pay per successful image: 1k and 2k have one shared rate, and 4k has three times that rate. Failed generations cost nothing.
| Specification | Price |
|---|---|
| $0.036 / image |
| $0.036 / image |
| $0.108 / image |
Use cases
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.
Prompt guide & pro tips
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.
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.
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.
Inspect the returned details
Check quoted lettering against the prompt and compare edited faces or objects with their sources. Save the chosen result and its settings before changing one part of the instruction for another run.
Generate a still from a written scene
MiniMax H3-Image text-to-image creates a picture without a source upload. Use it for an adult editorial portrait, a product scene or a composition containing a short message. The scene comes from your English description; canvas shape, pixel tier and encoding are selected through separate parameters. The default request produces a square 1k PNG.
Parameters for this endpoint
| Parameter | Accepted value and behavior |
|---|---|
prompt | Required English scene description, up to 5,000 characters |
aspect_ratio | Defaults to 1:1; supports the fifteen ratios listed below |
resolution | 1k by default, or 2k or 4k; changes the generation tier |
output_format | png by default; jpeg and webp are alternatives without a format surcharge |
seed | Optional integer from 0 to 2,147,483,647; repeats a starting point only while prompt expansion is off |
enable_prompt_expansion | true by default, so MiniMax rewrites the description into a fuller English prompt; false keeps your wording |
width, height | Read only with aspect_ratio set to custom; 468–7,745 pixels each, with the area inside the chosen tier |
Supported ratios are 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 and 21:9. A tall portrait and a wide room can use the same resolution tier with different dimensions. Set the intended shape in aspect_ratio so the framing instruction and canvas agree. Pick custom when the delivery slot needs an exact pixel size, such as 1920 x 1080; at 1k or 2k the area can reach 4.5 megapixels, and 4k covers larger frames up to 20 megapixels.
Pixel tiers and billing
The generation charge is per returned image. The 1k and 2k tiers share a rate, despite covering about one and four megapixels respectively. Selecting 4k raises the generation charge to three times that shared rate. JPEG, PNG and WebP have the same pricing rule. Consult the current rates on the MiniMax H3-Image model page before running a request; failed generations are not charged.
Our square measurements were 1024 x 1024 at 1k and 4096 x 4096 at 4k. At 16:9, the 2k result measured 2720 x 1536 and the 4k result measured 5472 x 3072. Treat the tier as an approximate area class rather than a promise that every image has a 4096-pixel edge.
Build a prompt around visible decisions
MiniMax says H3-Image expects a detailed English description of the whole picture. Start with who or what appears, then connect the wardrobe or material, setting, mood and illumination in ordinary sentences. For people, specify adults in their thirties or older. A grey-bearded man by a rainy window and an amber bottle on dark stone need different direction; give each scene the details that determine its appearance.
For lettering, put the exact phrase in quotation marks and attach it to an object. For example: a handwritten card reading "Room 1207, midnight", held toward the camera in soft morning light. That is a short, inspectable requirement. Check the returned words against your prompt rather than assuming that readable text in one sample establishes paragraph-length typography.
A useful revision sequence is:
- Save the prompt, selected ratio, tier and returned image together.
- Inspect the subject, requested wording and arrangement separately.
- Revise a concrete detail, such as the light direction or label placement.
- Compare the next result with the accepted direction before making another change.
Prompt expansion is on unless you set enable_prompt_expansion to false. While it is on, MiniMax rewrites the description before drawing, so a repeated seed can still produce a different picture. For a controlled comparison, switch expansion off, keep seed fixed and change one phrase at a time; the seed supports that comparison without promising an identical image. Negative prompt, guidance and sampling steps are not offered, so keep visual requirements in the main description.
What our samples and tests showed
The new Room 1207 sample used 2k, 3:2 and PNG. It returned a woman in a black silk slip across hotel sheets, holding a legible note with the requested "Room 1207, midnight" wording. Separate tests rendered the bottle label NOIR 07 clearly and showed a readable MIDNIGHT JAZZ sign. These observations establish short messages, labels and signage in those scenes; long passages remain unverified.
A six-prompt comparison at 1k also reproduced "HOT SOUP 3 FRANCS" on a wooden sign across H3-Image, Open Weights and P-H3-Image. The matching phrase did not make the pictures equivalent. H3-Image placed more small Latin labels on a botanical plate and followed a requested pedestrian-bridge viewpoint in that comparison. Those are individual observations, without a general quality ranking.
Measured model processing on November 11, 2026 was about 12–15 seconds at 1k, with 4k observations around 20–22 seconds. Waiting in a queue and transferring the file add time beyond those measurements.
Content behavior has variation
No request was refused in our standardized NSFW tests, and every delivered image was billed. At Very strong, text-to-image matched the requested result in 1/3 runs; at Maximum it matched 6/6. Named levels describe the test record without guaranteeing the outcome of another request. A toned-down delivered picture is distinct from a failed generation for billing purposes.
Choose a related workflow when needed
Move to image-to-image on this page when a source face, object or environment must guide the composition. Open Weights runs the released weights with five size tiers instead of three. Use Open Weights LoRA when the requested workflow includes LoRA files. These alternatives have their own controls and pricing, and matching output across versions has not been verified.
API quick start
# Generate the idempotency key once; reuse it when retrying after a lost response.
export SPICY_IDEMPOTENCY_KEY="${SPICY_IDEMPOTENCY_KEY:-$(uuidgen)}"
curl -X POST https://api.spicyapi.ai/api/v1/jobs/createTask \
-H "Authorization: Bearer $SPICY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: $SPICY_IDEMPOTENCY_KEY" \
-d '{
"model": "minimax/h3-image/text-to-image",
"input": {
"prompt": "A ceramic teapot on a linen cloth, soft window light, muted palette",
"resolution": "1k",
"aspect_ratio": "1:1",
"output_format": "png"
}
}'
# Copy data.taskId from the response into TASK_ID, then poll the status.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.spicyapi.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
-H "Authorization: Bearer $SPICY_API_KEY"
See the API reference for authentication, the complete schema, polling, and error handling.




























