Available now

Qwen Image 2.1 Pro: image editing

By Alibaba

Use Qwen Image 2.1 Pro to revise a reference image. Write what to change, and use the negative prompt to leave out details that would distract from the subject.

Model ID
Maker
Alibaba
Added
October 12, 2026

Model id

Use this in every request
qwen-image-2.1-pro/edit

This model has other modes with their own ids: qwen-image-2.1-pro.

Run it

Quote the input, start the generation with the quote, then poll the job. Set RAYWAKE_API_KEY first; see Authentication.

# 1. Price the exact input (quotes are valid for up to 10 minutes)
curl -s https://api.raywake.com/v1/quotes \
  -H "Authorization: Bearer $RAYWAKE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "qwen-image-2.1-pro/edit",
  "input": {
    "prompt": "Replace the white backdrop with pale blue. Keep the product, the printed label and the existing shadow.",
    "image_urls": [
      "https://example.com/input.png"
    ]
  }
}'

# 2. Start the generation with the quote_id from step 1
curl -s https://api.raywake.com/v1/generate \
  -H "Authorization: Bearer $RAYWAKE_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{"model": "qwen-image-2.1-pro/edit", "quote_id": "QUOTE_ID", "input": {"prompt":"Replace the white backdrop with pale blue. Keep the product, the printed label and the existing shadow.","image_urls":["https://example.com/input.png"]}}'

# 3. Poll until the status is terminal
curl -s https://api.raywake.com/v1/jobs/JOB_ID \
  -H "Authorization: Bearer $RAYWAKE_API_KEY"

Input

Input fields for qwen-image-2.1-pro/edit
NameTypeDescription
promptrequiredstring
Describe the edits to apply to the reference images in Chinese or English. Recommended maximum: 4,500 tokens.
enable_thinkingboolean
Enhance reasoning and image quality; takes longer. Effective only with prompt_extend enabled.

Default: true

image_urlsrequiredarray of image URLs
Provide 1 to 10 reference images in order. JPG, JPEG, PNG, BMP, TIFF, WEBP or GIF; up to 10 MB each. Recommended dimensions: 384–2048 pixels.
ninteger
Choose 1 to 6 output images.

Default: 1

negative_promptstring
Describe content to exclude from the generated image.
prompt_extendboolean
Enable LLM-based prompt rewriting for better results on simple prompts. Default: true.

Default: true

prompt_extend_modestring (enum)
Direct prompt enhancement only. Agent enhancement is not supported for image editing.

Default: direct

One of: direct

seedinteger
Random seed for reproducible generation. Range: 0-2147483647.
sizestring (enum)
Choose a preset output size.

Default: 1024*1024

One of: 512*512, 1024*1024, 2048*2048, 2688*1536, 1536*2688, 2368*1728, 1728*2368

Output

A succeeded job lists its results under outputs, stored on our side. See Jobs for every field.

Job excerpt
"outputs": [
  { "type": "image", "url": "https://…", "mime_type": "image/png", "thumbnail_url": "https://…" }
]

Pricing

  • per image$0.07/ image

The exact charge for your input comes from a quote. How charges settle

API docs

About this model

Overview

The generation form gives you preset canvas sizes and a transparent background option. The edit form takes reference images and follows a separate instruction.

What you can build

  • Transparent graphics for a layout
  • Product compositions from written briefs
  • Edits with details carried over from a reference

Prompting tips

  • Describe the framing and light before adding secondary props.
  • For an edit, name the object to change and the details to retain.
  • For a cutout, turn on Transparent background: prompt rewriting is switched off for that request and the price stays the same.

What to check

Inspect edges, hands, lettering and shadows at full size. For a transparent PNG, view it on a light and a dark background: Pro can keep a soft, semi-transparent shadow under the subject. Keep the prompt, mode, size and seed with the chosen image.

Inputs

  • Prompt (required): Describe the edits to apply to the reference images in Chinese or English. Recommended maximum: 4,500 tokens.
  • Enable Thinking (optional): Enhance reasoning and image quality; takes longer.
  • Images (required): Provide 1 to 10 reference images in order. JPG, JPEG, PNG, BMP, TIFF, WEBP or GIF; up to 10 MB each. Recommended dimensions: 384–2048 pixels.
  • Number of Images (optional): Choose 1 to 6 output images.
  • Negative Prompt (optional): Describe content to exclude from the generated image.
  • Prompt Extend (optional): Enable LLM-based prompt rewriting for better results on simple prompts. Default: true.
  • Prompt Enhancement Mode (optional): Direct prompt enhancement only. Agent enhancement is not supported for image editing. Options: direct.
  • Seed (optional): Random seed for reproducible generation. Range: 0-2147483647.
  • Size (optional): Choose a preset output size. Options: 512*512, 1024*1024, 2048*2048, 2688*1536, 1536*2688, 2368*1728, 1728*2368.

Output and limits

  • Small or dense text in the image can come out misspelled.
  • Hands, crowds and complex poses can show artifacts; generate a few variations and pick the best.
  • Some prompts or outputs can be refused by safety checks.

Price guide

  • per image: $0.07 / image