Image generation and editing

MiniMax H3 Image API

Create images, edit references, and refine your prompts.

Use Standard or Highspeed for text-to-image and reference-image editing. Choose 1K, 2K, 4K, or custom dimensions and collect images through the unified task API. Standalone prompt enhancement returns revised text for your next step.

Read the official model and pricing documentation

Try playground

1–10

Images per request

Up to 16

Reference images

Actual tokens

Billing

MiniMax

minimax-h3-imageimage task

1–10

Images per request

Up to 16

Reference images

Actual tokens

Billing

Model capabilities

Built for production API workflows, not one-off demos.

Use MiniMax H3 Image from the same platform surface as the rest of your video and image stack: API keys, credits, logs, webhooks, and docs stay consistent across providers.

Standard and Highspeed

Both versions support text-to-image, reference-image editing, custom dimensions, JPEG, PNG, and WebP output.

Prompt enhancement

Preview revised prompts independently. Text-to-image can disable built-in enhancement when the output ratio or size is explicit; editing always enhances prompts.

Composition controls

Choose automatic composition or custom aspect ratios from 1:3 to 3:1, with up to 20 million output pixels.

API workflow

Submit tasks, track progress, and return generated assets.

Describe or upload references

Write a prompt up to 32,000 Unicode characters. Add 1–16 public reference URLs to edit images.

Submit and track

Receive a task ID immediately. Query progress or use a completion callback. Submitted generation is never retried automatically.

Receive results and actual billing

Collect successful outputs and token usage. The task is charged once on completion, with no pre-charge.

POST /v1/images/generations

curl -X POST https://api.aivideoapi.ai/v1/images/generations \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{"model":"minimax-h3-image","input":{"prompt":"A quiet reading room in soft morning light","area":"2K","aspect_ratio":"16:9","n":1}}'

What teams build

Product imagery

Reference-guided edits

Campaign variations

Prompt refinement

Frequently asked questions

Answers about the MiniMax H3 Image API.

Is there a fixed price per image?

No. Actual text input, image input, and image output tokens determine the charge. Submit with a positive balance; the final debit may make the balance negative, in which case top up before starting another task.

How does the 4K offer work?

For requests dispatched before November 18, 2026, 00:00 Beijing time, every successful output must be at least 16.6 million pixels to receive 50% off the whole request. Mixed or missing sizes require billing review. Standalone enhancement is excluded.

What happens after a timeout?

A timeout can leave the result unknown. The same request is not generated again automatically. Stored successful results are retained for delivery and billing review; no final charge is guessed.

Pricing and usage

Clear model options with shared credits.

Standard · per million tokens

¥19.50 / ¥13.00 / ¥32.50

Text input / image input / image output

Highspeed · per million tokens

¥19.50 / ¥26.00 / ¥65.00

Text input / image input / image output

Standalone prompt enhancement

¥19.50 / million input tokens

CNY pricing; credits use 6.67 CNY/USD and USD 0.005/credit. Account multipliers apply.

Start building with MiniMax H3 Image in AI Video API.

Create one API key, use one credit balance, and switch between video and image models without provider-specific plumbing.

Read docs