MiniMax H3 Image. Generate and edit from many references.
MiniMax's image model makes a picture from a prompt, or edits and composes one from several reference images. It comes as H3 Image and H3 Image Highspeed, at 1K, 2K or 4K, in 23 aspect ratios.
- Text and image input
- 1K, 2K or 4K output
- 23 aspect ratios plus auto
- Up to 10 reference images per node
MiniMax H3 Image is one model for generating and editing stills. Write a prompt to start from nothing, or attach images and describe what should change: a product moved to a new set, a style carried from one picture to another, several references composed into one frame. It ships in two versions with the same inputs and settings, H3 Image and H3 Image Highspeed, which MiniMax lists as the faster one.
On Treza both sit in the Image Generation node's model picker. Pick 1K, 2K or 4K and one of 23 aspect ratios, from 1:3 portrait strips to 21:9 widescreen, or leave the ratio on auto. In our tests a 1K image came back in 10 to 13 seconds, and a finished still wires straight into the rest of the pipeline, which publishes as one API.
Up to 10 references
Attach products, people, settings and style frames to one node, and H3 composes a single image from them.
Generate and edit in one model
Start from a prompt, or attach a photo and describe the change. The same node does both.
1K, 2K or 4K
Draft at 1K, then render the keeper at 2K or 4K. Left unset, it renders 4K.
23 aspect ratios
From 1:3 and 9:21 tall strips to 21:9 and 2.35:1 widescreen, with auto when the prompt should decide.
What to make with MiniMax H3 Image
Prompts to try, and what to expect back. Each runs on an Image Generation node with H3 Image or H3 Image Highspeed picked in the model menu.
“A ceramic cup of coffee on a sunlit wooden cafe table, steam rising, shallow depth of field”
A warm, natural-light product still.
“The same folded shirt (attached) on a white marble counter in a bright studio”
The shirt, its weave and collar kept, on a new set.
“Put this sneaker (attached) on this street (attached), in the style of this frame (attached)”
One composition from three references.
“A 21:9 banner of a mountain lake at dawn for a website header”
A wide strip sized for the header.
“A 9:16 poster of a jazz night in a small club, moody blue light”
A vertical frame ready for a story or a short.
“The same jar (attached) in four seasonal settings, one per image”
A set of matching product shots for a campaign.
MiniMax H3 Image is one node on a canvas
Chain it with language, image, and video models, then publish the whole pipeline as an API. See how a full AI video pipeline fits together.
- Step 01
Draft the brief
A language model node expands a one-line idea into a detailed, on-brand prompt before it ever reaches the model.
- Step 02
Generate with MiniMax H3 Image
Run variants in parallel, compare against other models on the same prompt, and keep every result in run history.
- Step 03
Publish it as an API
The pipeline becomes a versioned endpoint. Your product calls it with the typed /invoke API or any OpenAI SDK.
MiniMax H3 Image as an API, priced per image
The MiniMax H3 Image API
Build a pipeline with MiniMax H3 Image on a node and publish it. It becomes a versioned HTTP endpoint: the typed /invoke route enqueues a run and returns a run id to poll, and an OpenAI-compatible route works with SDKs you already use. Agents connect over MCP the same way. Your prompt engineering ships inside the pipeline, so every caller gets your quality, and swapping the model later does not change the API your product calls.
MiniMax H3 Image pricing
Prepaid credits, no subscription. Each image is charged at MiniMax H3 Image's own rate from your balance, shown before you run, and a failed generation costs nothing. Credit packs start at $5 and never expire, which suits spiky production schedules. Run history records the cost of every generation, per node, so spend is auditable rather than a surprise.
MiniMax H3 Image, answered
What is MiniMax H3 Image?
MiniMax's image generation and editing model. It makes images from text prompts, and edits or composes images from reference images (up to 10 per node on Treza), at 1K, 2K or 4K in 23 aspect ratios.
What is the difference between H3 Image and H3 Image Highspeed?
Same inputs, settings and limits. MiniMax lists Highspeed as the faster version, and it is billed at about twice the rate per image. In our tests both returned a 1K image in 10 to 13 seconds, so run one prompt through both on two nodes and let the output decide.
What does MiniMax H3 Image cost?
It is billed by tokens, not a flat price per image. In our tests a 1K or 2K image billed about 2 cents on H3 Image and about 4 cents on Highspeed, and a reference image added under a cent. Treza quotes every run before it starts, plus its credit markup, and a failed generation costs nothing.
Can I use MiniMax H3 Image in chat?
Yes. Ask for MiniMax H3 Image or H3 Image Highspeed by name in Treza chat, or pick either one in the Image Generation node on the canvas.
How much does MiniMax H3 Image cost on Treza?
Treza uses prepaid credits with no subscription. Each image is charged at MiniMax H3 Image's rate from your balance, and only successful generations are charged. Credit packs start at $5 and never expire.
Can I call MiniMax H3 Image through an API?
Yes. Build a pipeline that uses MiniMax H3 Image and publish it as a versioned HTTP endpoint. Call the typed /invoke API from your product, or point any OpenAI SDK at the OpenAI-compatible endpoint. If a better model ships later, swap it on the node without changing the API your product calls.
Your first video is one prompt away.
Generate your first video, image, or draft today. Prepaid credits, no subscription.