---
title: MiniMax H3 Image | Runware Docs
url: https://runware.ai/docs/models/minimax-h3-image
description: Text-to-image generation and multi-reference editing at up to 20 megapixels
---
# MiniMax H3 Image

![MiniMax H3 Image](https://assets.runware.ai/covers/minimax-h3-image.jpg)

MiniMax H3 Image is an open-weights image generation and editing model from MiniMax. It creates images from text or edits and combines up to 16 reference images, at 1K, 2K or 4K, or at a custom width and height up to 20 megapixels. It follows counts, positions and layouts closely and renders long, dense text, which suits product and marketing visuals, posters and infographics, concept art, and compositions that merge several subjects.

- **Status**: live
- **ID**: `minimax:h3@image`
- Try in Playground
- **Creator**: MiniMax
- **Release Date**: November 10, 2026
- **Zero Data Retention**: Yes
- **Capabilities**: Text to Image, Image to Image, Edit, Checkpoint

## Pricing

~$0.0311 typical run

**Rates** (per 1M tokens)

- Text input: `$2.30`
- Image input: `$1.50`
- Image output: `$3.80`

**4 example generations**

- [text-to-image](https://runware.ai/docs/models/minimax-h3-image/examples#example-championship-pigeon-racer-portrait): `$0.0176` (~16s)
- [text-to-image](https://runware.ai/docs/models/minimax-h3-image/examples#example-japanese-joinery-workshop-wall-chart): `$0.0189` (~16s)
- [reference-to-image](https://runware.ai/docs/models/minimax-h3-image/examples#example-planetarium-restoration-fundraising-panorama): `$0.0432` (~27s)
- [reference-to-image](https://runware.ai/docs/models/minimax-h3-image/examples#example-underground-mushroom-grower-editorial-portrait): `$0.0432` (~31s)

billed per token, so a request costs what the model produced · times vary, measured end to end · [How pricing works](https://runware.ai/docs/models-api/pricing)

## Compatibility & Validation

When `inputs.referenceImages` is provided, `settings.promptOptimizer` must be `true`.

---

`resolution` cannot be used with `width/height`.

---

`width` and `height` must be used together.

---

When none of `width` and `height` are provided, `settings.promptOptimizer` must be `true`.

## Notes

**Modes:** a prompt alone generates an image. Adding `inputs.referenceImages` switches to editing, where the prompt refers to each reference by its position (image 1, image 2) and can give each one a role, such as the product from one image and the color palette from another.

**Prompt optimizer:** `settings.promptOptimizer` rewrites the prompt into a detailed scene description before generating, and editing always uses it. With it off, the prompt is used as written and the model reads only about its first 4,096 tokens.

**Size:** with `resolution`, the model sets the width and height itself, following the references or a ratio it picks from the prompt. With no size at all, the output is 4K. An exact `width` and `height` is generated as given, never scaled or rounded.

**Multi-round editing:** each request is independent. To keep editing a result, send it back as a reference image with the next instruction. Every round re-renders the whole image and adds contrast and sharpening, so quality drops over a long chain.

**Pricing:** billed per token. An output of up to about 4 megapixels uses about 4,100 image tokens, so 1K and 2K cost about the same, and 4K uses about 16,400. The rewritten prompt counts as text input, and each reference image adds about 2,050 input tokens.

## Request Parameters

**API Options**

Platform-level options for task execution and delivery.

### [taskType](https://runware.ai/docs/models/minimax-h3-image#request-tasktype)

- **Type**: `string`
- **Required**: true
- **Value**: `imageInference`

Identifier for the type of task being performed.

### [taskUUID](https://runware.ai/docs/models/minimax-h3-image#request-taskuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID v4 identifier for tracking tasks and matching async responses. Must be unique per task.

### [outputType](https://runware.ai/docs/models/minimax-h3-image#request-outputtype)

- **Type**: `string`
- **Default**: `URL`

Image output type.

**Allowed values**: `URL` `base64Data` `dataURI`

### [outputFormat](https://runware.ai/docs/models/minimax-h3-image#request-outputformat)

- **Type**: `string`
- **Default**: `JPG`

Specifies the file format of the generated output. The available values depend on the task type and the specific model's capabilities.

- \`JPG\`: Best for photorealistic images with smaller file sizes (no transparency).
- \`PNG\`: Lossless compression, supports high quality and transparency (alpha channel).
- \`WEBP\`: Modern format providing superior compression and transparency support.

> [!NOTE]
> \*\*Transparency\*\*: If you are using features like background removal or LayerDiffuse that require transparency, you must select a format that supports an alpha channel (e.g., \`PNG\`, \`WEBP\`, \`TIFF\`). \`JPG\` does not support transparency.

**Allowed values**: `JPG` `PNG` `WEBP`

### [outputQuality](https://runware.ai/docs/models/minimax-h3-image#request-outputquality)

- **Type**: `integer`
- **Min**: `20`
- **Max**: `99`
- **Default**: `95`

Compression quality of the output. Higher values preserve quality but increase file size.

### [webhookURL](https://runware.ai/docs/models/minimax-h3-image#request-webhookurl)

- **Type**: `string`
- **Format**: `uri`

Webhook URL that receives JSON responses via HTTP POST when generation tasks complete. For batch requests with multiple results, each completed item triggers a separate webhook call as it becomes available.

**Learn more** (1 resource):

- [Webhooks](https://runware.ai/docs/models-api/webhooks) (platform)

### [deliveryMethod](https://runware.ai/docs/models/minimax-h3-image#request-deliverymethod)

- **Type**: `string`
- **Default**: `sync`

Determines how the API delivers task results.

**Allowed values**:

- `sync` Returns results directly in the API response. Over HTTP, a task still running after 120 seconds continues as an async task and comes back with status 'processing'. Poll for its results using getResponse.
- `async` Returns an immediate acknowledgment with the task UUID. Poll for results using getResponse.

**Learn more** (1 resource):

- [Task Polling](https://runware.ai/docs/models-api/task-polling) (platform)

### [uploadEndpoint](https://runware.ai/docs/models/minimax-h3-image#request-uploadendpoint)

- **Type**: `string`
- **Format**: `uri`

Specifies a URL where the generated content will be automatically uploaded using the HTTP PUT method. The raw binary data of the media file is sent directly as the request body. For secure uploads to cloud storage, use presigned URLs that include temporary authentication credentials.

**Common use cases:**

- **Cloud storage**: Upload directly to S3 buckets, Google Cloud Storage, or Azure Blob Storage using presigned URLs.
- **CDN integration**: Upload to content delivery networks for immediate distribution.

```text
// S3 presigned URL for secure upload
https://your-bucket.s3.amazonaws.com/generated/content.mp4?X-Amz-Signature=abc123&X-Amz-Expires=3600

// Google Cloud Storage presigned URL
https://storage.googleapis.com/your-bucket/content.jpg?X-Goog-Signature=xyz789

// Custom storage endpoint
https://storage.example.com/uploads/generated-image.jpg
```

The content data will be sent as the request body to the specified URL when generation is complete.

### [safety](https://runware.ai/docs/models/minimax-h3-image#request-safety)

- **Path**: `safety.checkContent`
- **Type**: `object (1 property)`

Content safety checking configuration for image generation.

#### [checkContent](https://runware.ai/docs/models/minimax-h3-image#request-safety-checkcontent)

- **Path**: `safety.checkContent`
- **Type**: `boolean`

Enable or disable content safety checking. Increases total generation time and adds a small charge on top of the generation cost.

### [ttl](https://runware.ai/docs/models/minimax-h3-image#request-ttl)

- **Type**: `integer`
- **Min**: `60`

Time-to-live (TTL) in seconds for generated content. Only applies when `outputType` is `URL`.

### [includeCost](https://runware.ai/docs/models/minimax-h3-image#request-includecost)

- **Type**: `boolean`

Include task cost in the response.

### [numberResults](https://runware.ai/docs/models/minimax-h3-image#request-numberresults)

- **Type**: `integer`
- **Min**: `1`
- **Max**: `20`
- **Default**: `1`

Number of results to generate. Each result uses a different seed, producing variations of the same parameters.

**Inputs**

Input resources for the task (images, audio, etc). These must be nested inside the \`inputs\` object.

### [referenceImages](https://runware.ai/docs/models/minimax-h3-image#request-inputs-referenceimages)

- **Path**: `inputs.referenceImages`
- **Type**: `array of strings`
- **Min items**: `1`
- **Max items**: `16`
- **Size (each)**: `128 px to 8192 px`
- **Aspect ratio (each)**: `1:4 to 4:1`
- **File size (each)**: `up to 64 MB`

List of reference images (UUID, URL, Data URI, or Base64).

**Core Parameters**

Primary parameters that define the task output.

### [model](https://runware.ai/docs/models/minimax-h3-image#request-model)

- **Type**: `string`
- **Required**: true
- **Value**: `minimax:h3@image`

Send `auto` instead of an AIR identifier to have the model router choose the model for this task. The rest of the request is unchanged, every value you send is kept, and only the parameters you leave out are filled in. See [Model Router](https://runware.ai/docs/models-api/model-router).

**Learn more** (3 resources):

- [Text To Image: Model Selection](https://runware.ai/docs/learn/text-to-image#model-selection) (learn)
- [Image Inpainting: Model Specialized Inpainting Models](https://runware.ai/docs/learn/image-inpainting#model-specialized-inpainting-models) (learn)
- [Image Outpainting: Other Critical Parameters](https://runware.ai/docs/learn/image-outpainting#other-critical-parameters) (learn)

### [positivePrompt](https://runware.ai/docs/models/minimax-h3-image#request-positiveprompt)

- **Type**: `string`
- **Required**: true
- **Min**: `1`
- **Max**: `32000`

Text prompt describing elements to include in the generated output.

**Learn more** (1 resource):

- [Prompts](https://runware.ai/docs/learn/prompts) (learn)

### [width](https://runware.ai/docs/models/minimax-h3-image#request-width)

- **Type**: `integer`
- **Min**: `468`
- **Max**: `7745`

Width of the generated media in pixels.

**Learn more** (2 resources):

- [Dimensions](https://runware.ai/docs/learn/dimensions) (learn)
- [Image Outpainting: Dimensions Critical For Outpainting](https://runware.ai/docs/learn/image-outpainting#dimensions-critical-for-outpainting) (learn)

### [height](https://runware.ai/docs/models/minimax-h3-image#request-height)

- **Type**: `integer`
- **Min**: `468`
- **Max**: `7745`

Height of the generated media in pixels.

**Learn more** (2 resources):

- [Dimensions](https://runware.ai/docs/learn/dimensions) (learn)
- [Image Outpainting: Dimensions Critical For Outpainting](https://runware.ai/docs/learn/image-outpainting#dimensions-critical-for-outpainting) (learn)

### [resolution](https://runware.ai/docs/models/minimax-h3-image#request-resolution)

- **Type**: `string`

Resolution preset for the output. When used with input media, automatically matches the aspect ratio from the input. Without reference images, the model picks the aspect ratio from the prompt.

**Allowed values**: `1K` `2K` `4K`

**Settings**

Technical parameters to fine-tune the inference process. These must be nested inside the \`settings\` object.

### [promptOptimizer](https://runware.ai/docs/models/minimax-h3-image#request-settings-promptoptimizer)

- **Path**: `settings.promptOptimizer`
- **Type**: `boolean`
- **Default**: `true`

Automatic prompt optimization. Affects reproducibility.

**Features**

Standalone addons and post-processing features.

### [watermark](https://runware.ai/docs/models/minimax-h3-image#request-advancedfeatures-watermark)

- **Path**: `advancedFeatures.watermark`
- **Type**: `object (6 properties)`

Configuration object for adding watermarks to generated videos. Watermarks can be applied using either text or image content with customizable positioning and appearance. You must provide either `text` or `image` content for the watermark, but not both.

**Text watermark**:

```json
"advancedFeatures": {
  "watermark": {
    "text": "© 2025 Company",
    "displayPosition": "bottom-right",
    "opacity": 0.6,
    "fontColor": "#ffffff",
    "bgColor": "#000000"
  }
}
```

**Image watermark**:

```json
"advancedFeatures": {
  "watermark": {
    "image": "c64351d5-4c59-42f7-95e1-eace013eddab",
    "displayPosition": "top-left",
    "opacity": 0.6
  }
}
```

**Tiled watermark**:

```json
"advancedFeatures": {
  "watermark": {
    "text": "PREVIEW",
    "tiled": true,
    "opacity": 0.4,
    "fontColor": "#cccccc"
  }
}
```

#### [text](https://runware.ai/docs/models/minimax-h3-image#request-advancedfeatures-watermark-text)

- **Path**: `advancedFeatures.watermark.text`
- **Type**: `string`
- **Min**: `2`
- **Max**: `32`

Watermark text.

#### [image](https://runware.ai/docs/models/minimax-h3-image#request-advancedfeatures-watermark-image)

- **Path**: `advancedFeatures.watermark.image`
- **Type**: `string`

Watermark image (UUID, URL, Data URI, or Base64).

#### [displayPosition](https://runware.ai/docs/models/minimax-h3-image#request-advancedfeatures-watermark-displayposition)

- **Path**: `advancedFeatures.watermark.displayPosition`
- **Type**: `string`

Watermark position.

**Allowed values**: `top-left` `top-center` `top-right` `center-left` `center-center` `center-right` `bottom-left` `bottom-center` `bottom-right` `tiled`

#### [opacity](https://runware.ai/docs/models/minimax-h3-image#request-advancedfeatures-watermark-opacity)

- **Path**: `advancedFeatures.watermark.opacity`
- **Type**: `float`
- **Min**: `0.1`
- **Max**: `1`
- **Step**: `0.01`

Watermark opacity.

#### [fontColor](https://runware.ai/docs/models/minimax-h3-image#request-advancedfeatures-watermark-fontcolor)

- **Path**: `advancedFeatures.watermark.fontColor`
- **Type**: `string`

Text color in hex format.

#### [bgColor](https://runware.ai/docs/models/minimax-h3-image#request-advancedfeatures-watermark-bgcolor)

- **Path**: `advancedFeatures.watermark.bgColor`
- **Type**: `string`

Background color in hex format.

## Response Parameters

### [taskType](https://runware.ai/docs/models/minimax-h3-image#response-tasktype)

- **Type**: `string`
- **Required**: true
- **Value**: `imageInference`

Identifier for the type of task this response belongs to.

### [taskUUID](https://runware.ai/docs/models/minimax-h3-image#response-taskuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID v4 identifier echoed from the original request, used to match async responses to their tasks.

### [imageUUID](https://runware.ai/docs/models/minimax-h3-image#response-imageuuid)

- **Type**: `string`
- **Required**: true
- **Format**: `UUID v4`

UUID of the output image.

### [imageURL](https://runware.ai/docs/models/minimax-h3-image#response-imageurl)

- **Type**: `string`
- **Format**: `uri`

URL of the output image.

### [imageBase64Data](https://runware.ai/docs/models/minimax-h3-image#response-imagebase64data)

- **Type**: `string`

Base64-encoded image data.

### [imageDataURI](https://runware.ai/docs/models/minimax-h3-image#response-imagedatauri)

- **Type**: `string`
- **Format**: `uri`

Data URI of the output image.

### [NSFWContent](https://runware.ai/docs/models/minimax-h3-image#response-nsfwcontent)

- **Type**: `boolean`

Flag indicating if NSFW content was detected.

### [cost](https://runware.ai/docs/models/minimax-h3-image#response-cost)

- **Type**: `float`

Task cost in USD. Present when `includeCost` is set to `true` in the request.

## Examples

### text-to-image (text-to-image)

![Japanese Joinery Workshop Wall Chart](https://assets.runware.ai/examples/minimax-h3-image/b8c2c31d-1d15-4f71-b68b-fa70ea8c71be.jpg)

```json
{
  "taskType": "imageInference",
  "taskUUID": "bf3b18ca-5af9-48ec-8b56-e585f9c0e82f",
  "model": "minimax:h3@image",
  "positivePrompt": "Create a finished landscape wall-art print, not a framed mockup and not a photograph of a poster. Design a clear educational chart about traditional Japanese timber joinery. Use a strict 5:4 landscape composition with a wide title band across the top, six joint diagrams arranged in an exact 3-column by 2-row grid, and a narrow footer band. Leave generous margins and consistent spacing.\n\nAt the top center, set the main title exactly as \"THE QUIET LOGIC OF WOOD\" in large uppercase lettering. Directly beneath it, set the subtitle exactly as \"SIX JAPANESE JOINTS — NO NAILS, NO GLUE\" in smaller uppercase lettering.\n\nShow exactly six distinct exploded isometric timber-joint diagrams, one per grid cell. Each diagram consists of two warm cedar components separated slightly to reveal how they interlock, with crisp black construction lines and small vermilion-red directional arrows. Number the diagrams with large red numerals in their upper-left corners and place two centered label lines beneath each illustration. Arrange and label them exactly as follows:\n\nTop left: \"01\" with \"KANAWA TSUGI\" and beneath it \"SCARF JOINT\".\nTop center: \"02\" with \"ARI TSUGI\" and beneath it \"DOVETAIL SPLICE\".\nTop right: \"03\" with \"KAMA TSUGI\" and beneath it \"GOOSENECK SPLICE\".\nBottom left: \"04\" with \"HOZO SASHI\" and beneath it \"MORTISE & TENON\".\nBottom center: \"05\" with \"SHACHI SEN\" and beneath it \"LOCKING KEY\".\nBottom right: \"06\" with \"OKKAKE DAISEN TSUGI\" and beneath it \"RABBETED SCARF\".\n\nIn the bottom center footer, write exactly \"MEASURE TWICE • CUT ONCE • FIT BY HAND\". Include no other words, letters, captions, logos, signatures, or watermarks.\n\nArt direction: refined Japanese technical print combining an Edo-period woodblock sensibility with a mid-century workshop diagram. Deep indigo background, warm honey-cedar wood grain, aged ivory typography, restrained vermilion accents, subtle fibrous washi-paper texture, slightly imperfect ink edges, clean geometric draftsmanship, elegant negative space, balanced visual hierarchy. The joints must remain mechanically plausible and easy to compare. Flat archival print appearance with gentle ink variation rather than glossy digital effects. Calm, disciplined, tactile, and suitable for premium print-on-demand wall art in a woodworker's studio.",
  "width": 1280,
  "height": 1024
}
```

---

### text-to-image (text-to-image)

![Championship Pigeon Racer Portrait](https://assets.runware.ai/examples/minimax-h3-image/9b3cd81e-7ea4-481c-973c-dbcf5347ee30.jpg)

```json
{
  "taskType": "imageInference",
  "taskUUID": "92fe548f-683c-48c7-a17c-08e4c15c1873",
  "model": "minimax:h3@image",
  "positivePrompt": "Create a finished landscape editorial photoshoot portrait of one championship pigeon racer on the rooftop of her urban pigeon loft. Compose her three-quarter length in the center-right of the frame, looking directly into the camera with quiet confidence. She is a stylish British-Caribbean woman in her late 50s with short silver curls, a weathered burgundy waxed field jacket, mustard silk blouse, tailored charcoal trousers, fine gold hoop earrings, and a vintage brass racing timer hanging from a narrow leather strap. Her left forearm is raised in a thick falconry-style glove.\n\nInclude exactly six pigeons in the entire image, all fully visible and clearly separated: one blue-bar racing pigeon perched on her raised glove; two pigeons spaced apart on the rooftop railing at far left; and three pigeons spaced apart along the pitched roof of the wooden loft at upper right. No other birds anywhere, including in the sky, shadows, reflections, background, or decorative imagery. The loft has small geometric entry openings, but no birds inside them. Keep the woman's face and all six pigeons sharply legible.\n\nFrame the scene as a sophisticated personal profile photograph with useful negative space in the upper-left sky for optional editorial cropping, but include no typography, captions, logos, borders, or watermarks. Show a dense London roofline softly out of focus behind her, with chimney pots, old brick, and distant apartment towers. Use a low eye-level camera angle and natural 55mm portrait perspective, realistic anatomy, authentic fabric texture, detailed feathers, and restrained depth of field.\n\nArt direction: contemporary British sporting portrait with subtle 1970s editorial character. Palette of burgundy, mustard, weathered teal, slate blue, warm brick, and silver gray. Late-afternoon sunlight breaks through after rain, creating a warm rim light around her hair and the pigeons while cool storm clouds remain in the distance. Slightly damp rooftop surfaces catch soft reflections. Mood is dignified, resilient, intimate, and quietly triumphant rather than whimsical. Photorealistic high-end magazine photography, fine natural grain, nuanced skin texture, controlled contrast, polished color grade.",
  "width": 1280,
  "height": 1024
}
```

---

### reference-to-image (reference-to-image)

![Planetarium Restoration Fundraising Panorama](https://assets.runware.ai/examples/minimax-h3-image/2dd58ae8-c210-4970-99e2-0f2cd23da4a9.jpg)

```json
{
  "taskType": "imageInference",
  "taskUUID": "7b15f3c1-0d13-47cf-9bca-b070e8a0632b",
  "model": "minimax:h3@image",
  "positivePrompt": "Create a seamless 16:9 outpaint for a museum fundraising campaign banner. Use image 1 as the immutable central source photograph: preserve its antique Zeiss star projector, conservator, pose, clothing, scaffold, tools, lighting, perspective, proportions, surface details and central framing unchanged. Do not redesign, duplicate, move or crop the projector or conservator. Extend the photograph naturally beyond image 1 on the left and right, with no visible borders, seams or change in photographic texture. Use image 2 as the architectural and color reference for the newly revealed surroundings: continue the ribbed plaster dome, maintenance catwalk, curved rows of faded oxblood seats, dark blue walls, amber aisle lamps and authentic age patina. Keep image 2's distinctive ceiling-rib geometry, seat color, indigo-brass-burgundy palette and historic material character unchanged while adapting its perspective to connect convincingly with image 1. The expanded sides should reveal more of the auditorium without adding another person or another projector. Maintain consistent lens perspective, grain, depth of field, shadow direction and practical work-light illumination across the complete panorama. Compose with the original restoration activity near the center and generous environmental context on both sides, suitable for a 2K museum campaign hero image. Cinematic documentary realism, restrained and dignified mood, historically credible conservation workspace. No lettering, logo, border, collage, split screen or interface.",
  "resolution": "2K",
  "inputs": {
    "referenceImages": [
      "https://assets.runware.ai/assets/inputs/d48e3551-b387-473f-ad03-5608212b8a16.jpg",
      "https://assets.runware.ai/assets/inputs/d3dc5009-b67e-4d2f-b216-f16b18df61c8.jpg"
    ]
  }
}
```

---

### reference-to-image (reference-to-image)

![Underground Mushroom Grower Editorial Portrait](https://assets.runware.ai/examples/minimax-h3-image/6cf79810-aaa8-4846-86d9-788d61c8b3e5.jpg)

```json
{
  "taskType": "imageInference",
  "taskUUID": "8247ed34-c1e6-4d2b-babf-5c97a8c3cc74",
  "model": "minimax:h3@image",
  "positivePrompt": "Create a finished horizontal editorial magazine photograph for the opening spread of a serious feature about subterranean urban agriculture. Use image 1 as the subject reference: place the mushroom grower from image 1 inside the farm, preserving her facial identity, age, skin tone, short black bob, cream utility coveralls, dark green rubber gloves, and black work boots unchanged. She stands slightly right of center beside a waist-high rack, one gloved hand gently supporting a cluster of oyster mushrooms while looking directly into the camera with quiet confidence. Use image 2 as the location and visual palette reference: preserve the tunnel's damp brick arches, galvanized rack geometry, oyster mushroom grow bags, fluorescent fixtures, wet concrete floor, cool blue-green ambient color, and deep receding perspective unchanged. Integrate the subject naturally at the correct scale with convincing contact shadows and reflected light. Photograph from waist height with a 40mm medium-format documentary look, environmental portrait framing, crisp eyes and tactile detail in the mushrooms, brick, condensation, fabric, and gloves. Mix the existing cool fluorescent ambience with restrained direct on-camera flash for a contemporary editorial aesthetic: honest, slightly stark, sophisticated, and unsentimental. Keep generous darker negative space in the upper-left area for an editor to place a headline later, but generate no typography, captions, logos, borders, watermark, or magazine mockup. Photorealistic color photography, natural anatomy, realistic skin texture, subtle film grain, controlled highlights, 2K-ready detail.",
  "resolution": "2K",
  "inputs": {
    "referenceImages": [
      "https://assets.runware.ai/assets/inputs/268e97ce-36ec-4227-9edf-9895fe68842f.jpg",
      "https://assets.runware.ai/assets/inputs/8296f3ce-8443-4662-8219-b0bf4e117640.jpg"
    ]
  }
}
```