logo
RunComfy
  • ComfyUI
  • Trainer
  • Models
  • MCPNew
  • Pricing
  • Seedance 2.5Hot
  • Seedream 5.5New

Workflows by Model

HotMiniMax H3Video with audio from text, images or referencesHotQwen Image 2.1Create images and edit photos with referencesWanVideo from text or images, motion transferFLUXImage generation, editing and character variationsLTXVideo with synced audio from text or imagesQwen Image EditPhoto editing with text and image referencesKrea 2Image creation, style transfer and reference editsZ-ImageFast portraits and concepts from text

Popular & New Workflows

NewMiniMax H3 Character Swap V1.5Swap video characters, keep the original audioNewP-H3-Image Reference EditEdit a photo guided by a reference imageHotMiniMax H3 Image-to-VideoOne image to cinematic video with synced audioHotLTX 2.5 Image-to-VideoOne frame to sharp video with synced audioHotMiniMax H3 Character Swap (SAM3)Swap video characters while following the motionQwen Image 2.1 Multi-Image EditingPrecise scenes from up to 10 reference imagesFLUX Klein Face SwapNatural, lifelike face swapsSeedVR2 Video UpscalerSharp, smooth video upscaling
View More Workflows

Train a Model

NewMOSS-Image BaseImage generation, newly supportedNewFLUX.3 KleinImage generation, newly supportedHotKrea 2The most-trained image modelHotKrea 2 TurboFew-step image generationHotQwen Image 2.1Image generation and editingZ-Image TurboFast portraits and concept imagesMiniMax H3Video with synchronized audio

Train & Deploy

Getting StartedYour first training job with Ostris AI ToolkitDatasetsUpload and manage training datasetsInferenceRun your trained models in a playground or via APIDeploymentsServe your trained models as dedicated endpoints
Train More Models

Image Models

NewGPT Image 2.65% OFF · Detailed text to image, GIF and SVG, up to 4KNewSeedream 5.510% OFF · Text to image up to 2K, with transparent PNGsNewMiniMax H3 Image10% OFF · Text or references to a still up to 4KNewFLUX 3 Image Klein10% OFF · Text or references to a still from 768 to 4KHotSeedream 5.0 Pro41% OFF · Reference edits with multilingual textHotSeedream 4.520% OFF · Photoreal edits with brand consistencyHotWan 2.6 ImageInstruction-based image editingNano Banana Pro77% OFF · Precise 2K to 4K visuals from sketches

Video Models

NewSeedance 2.5 Flash34% OFF · Text, image, or reference to videoNewKling 4.010% OFF · Animate a still into a 3 to 30s clipHotWan 3.0 Prime13% OFF · Fast clips from reference videosHotSeedance 2.536% OFF · Longer, more consistent cinematic videoHotWan 3.0Clips from image, video and audio referencesHotMiniMax H3Animate a first frame into a 5 to 15s clipWan 2.510% OFF · Fluid motion with native audioFLUX 3 MaxClips from text, a still, or keyframes
View More Models

API

Serverless (ComfyUI)Deploy ComfyUI workflows as serverless endpoints.Serverless (LoRA)Deploy LoRAs as serverless endpoints.Model APIRun hosted models on-demand with no deployment.Trainer APITrain, manage, and retrieve LoRAs via API.CLIRun RunComfy models from your terminal.
MODELS
Explore
All Models
LIBRARY
Generations
MODEL APIS
API Docs
API Keys
ACCOUNT
Usage

MiniMax H3 Image: Text-to-Image up to 4K on Models and API | RunComfy

minimax/minimax-h3-image/text-to-image

MiniMax H3 Image turns a text prompt into one 1K, 2K, or 4K still. 1K and 2K cost the same.

Describe the whole frame: subject, place, light, and style. Chinese or English, up to 5000 characters. Quoted words are more likely to be drawn as text.
Frame shape. auto lets prompt expansion choose. With expansion off, auto falls back to 1:1.
1K is about 1 megapixel, 2K about 4, 4K about 16. 1K and 2K are the same price. 2K is the default.
Rewrite the prompt into a fuller English description before generating. Leave this on unless you need a repeatable seed.
PNG is the default. JPEG is smaller.
Use -1 for a new result each run. A fixed seed repeats only when Expand Prompt is off.
Idle
1K and 2K are $0.029 per image. 4K is $0.085 per image.

Introduction To MiniMax H3 Image

MiniMax's MiniMax H3 Image writes one still from a text prompt at 1K, 2K, or 4K, up to about 16 megapixels.

Trading a separate enlarge step for a native large frame, MiniMax H3 Image helps poster, product, and campaign teams start from a short brief.

For developers, MiniMax H3 Image on RunComfy can be used both in the browser and via an HTTP API, so you don't need to host or scale the model yourself.

Ideal for: Poster Key Art | Product Stills | Campaign Frames

Why Choose MiniMax H3 Image#


MiniMax H3 Image is the still-image pass of MiniMax's H3 family. One prompt becomes one picture. The 4K tier is drawn at about 16 megapixels, from a 4096 square to a wide 21:9 frame, without a second enlarge job.


MiniMax H3 Image offers three size tiers. 1K is about 1 megapixel. 2K is about 4 megapixels and is the default. 4K is about 16 megapixels. 1K and 2K share one price, so 2K is the usual keep. Ten shapes are available, plus auto. Sampling is fixed at six steps. There is no negative prompt and no step slider.


The edit page is separate. MiniMax H3 Image on this page does not take reference files. Use the edit page when the job starts from one to sixteen pictures.


Best Use Cases#


  • A poster plate: Use MiniMax H3 Image when the frame has to be large enough to print or crop.
  • A product still: Describe the object, the surface, and one light. Leave logos out of the sentence unless the words must appear.
  • A campaign crop: On MiniMax H3 Image, pick 9:16 or 21:9 when the same idea has to fit a story or a banner.
  • A readable line: Put the few words you need in quotes. The model can draw short English or Chinese text, and it still misses letters when the line is long.
  • A first look: Stay on 2K while you search. Move to 4K only for the file you will keep.

How It Works#


  1. Open MiniMax H3 Image on RunComfy. Use the edit page if you already have pictures to change.
  2. Write the whole frame: subject, place, light, and style.
  3. Leave Expand Prompt on. Turn it off only when you will reuse a seed.
  4. Choose a shape, or leave auto so the rewrite can pick one for MiniMax H3 Image.
  5. Keep resolution on 2K unless you need the smaller 1K file or the 4K file.
  6. Pick PNG or JPEG. PNG is the default.
  7. Run it. MiniMax H3 Image returns one image per request.
  8. Send the same prompt and settings through the HTTP API.

Parameters#


ParameterRequiredTypeDefaultRange / OptionsDescription
prompt*Yes (*)stringExample lineUp to 5000 charactersThe scene for MiniMax H3 Image. Chinese or English.
aspect_ratioNostringautoauto, 1:1, 5:4, 4:5, 4:3, 3:4, 3:2, 2:3, 16:9, 9:16, 21:9Frame shape.
resolutionNostring2K1K, 2K, 4KSize tier for MiniMax H3 Image.
expand_promptNobooleantruetrue, falseRewrite the prompt before the still is drawn.
output_formatNostringpngpng, jpegFile type.
seedNointeger-1-1 to 2147483647Repeat a result only with expansion off.

Pricing#


MiniMax H3 Image bills one output image. Aspect ratio, prompt expansion, seed, and file type do not add a fee.


ResolutionPrice per image
1K$0.029
2K$0.029
4K$0.085

A 2K still is the usual first charge on MiniMax H3 Image.


Prompting Tips#


  • Name the whole frame: MiniMax H3 Image does better with subject, place, and light in one paragraph than with a pile of tags.
  • One change at a time: If a rerun is wrong, rewrite one clause. The step count cannot be raised.
  • Quote short text: A few words in quotes are the practical limit. A paragraph of type usually breaks.
  • Skip the negative: There is no negative field. Say what you want in the frame instead.
  • Lock a seed last: Turn expansion off, then reuse the seed. With expansion on, the rewritten line changes.
  • Prefer 2K: It costs the same as 1K and holds more detail.

How MiniMax H3 Image Compares#


Based on publicly available information. These notes describe inputs, not a scored ranking.


ModelInputsSizeStandout
MiniMax H3 ImageText only on this page1K, 2K, or 4KNative large still, six fixed steps
MiniMax H3 Image EditOne to sixteen reference picturesSame tiersChanges or combines pictures you already have
Earlier MiniMax still modelsA shorter prompt and fewer size choicesSmaller published framesUseful when a workflow is already built on them

Stay on this page for a new still. Open the edit page when the source is a picture.


Official Resources#


  • MiniMax: https://www.minimax.io
  • Weights: https://huggingface.co/MiniMaxAI/MiniMax-H3-Image

Related Models

reve/edit

Transform visuals with smart region edits and multi-image blending for precise, high-fidelity results.

z-image/turbo/text-to-image

High-speed model for rapid text-to-image creation with rich detail and flexible format control.

flux-1-kontext/dev/image-to-image

Edit visuals via text with multi-layer control and style memory.

nano-banana-2.1/edit

Prompt-led image edits from up to 10 references, at 1K, 2K, or 4K.

minimax-h3-image/edit

Edit 1 to 16 reference images into one 1K, 2K, or 4K still.

flux-3/text-to-image

FLUX 3 Image generates detailed text-to-image pictures; FLUX 3 Image Edit covers local edits and references

Frequently Asked Questions

What is MiniMax H3 Image used for?

MiniMax H3 Image makes one still from a text prompt at 1K, 2K, or 4K. The 4K tier is about 16 megapixels and is drawn in one pass. Poster, product, and campaign teams use MiniMax H3 Image when the frame has to be large enough to crop.

How does MiniMax H3 Image handle text inside a picture?

MiniMax H3 Image can draw short English or Chinese wording when you put those words in quotes. Long lines and small type still break. Treat MiniMax H3 Image as reliable for a title of a few words, then check every letter.

Which size should I pick in MiniMax H3 Image?

1K is about 1 megapixel, 2K about 4, and 4K about 16. 1K and 2K cost the same on MiniMax H3 Image, and 2K is the default, so start there. Move to 4K only for the file you will keep.

What input limits should I know before using MiniMax H3 Image?

The prompt can be up to 5000 characters in Chinese or English. Shapes are auto, 1:1, 5:4, 4:5, 4:3, 3:4, 3:2, 2:3, 16:9, 9:16, and 21:9. MiniMax H3 Image uses six fixed steps, has no negative prompt, and returns one image per request.

Can I reproduce a MiniMax H3 Image result?

Set a seed and turn Expand Prompt off. With expansion on, the rewritten line changes between runs, so the same seed will not match. MiniMax H3 Image only repeats when the instruction sent to the model is unchanged.

How is MiniMax H3 Image different from the edit page?

This page takes text only. The edit page takes 1 to 16 reference images and is the place to change a picture you already have. MiniMax H3 Image here is for a new still.

Can developers call MiniMax H3 Image from the RunComfy API?

Yes. Try the prompt and settings in the RunComfy model UI, then send the same fields to the RunComfy API. MiniMax H3 Image uses one request per image. The API does not add a speed guarantee.

How much does MiniMax H3 Image cost on RunComfy?

MiniMax H3 Image is $0.029 per image at 1K or 2K, and $0.085 at 4K. Aspect ratio, prompt expansion, seed, and PNG versus JPEG do not change the price. Failed tasks are not billed.

Follow us
  • LinkedIn
  • Facebook
  • Instagram
  • Twitter
Support
  • Discord
  • Email
  • System Status
  • Affiliate
Video Models
  • Seedance 2.5 Flash Text To Video
  • Seedance 2.5 Flash
  • Seedance 2.5 Flash Reference to Video
  • Wan 3.0
  • MiniMax H3 Fast
  • Wan 2.5
  • View All Models →
Image Models
  • Seedream 5.5
  • Seedream 5.5 Edit
  • Seedream 5.5 Layer Decomposition
  • Wan 2.6 Image to Image
  • Nano Banana Pro
  • Seedream 5.0 Pro
  • View All Models →
Legal
  • Terms of Service
  • Privacy Policy
  • Cookie Policy
RunComfy
Copyright 2026 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.

Examples Of MiniMax H3 Image