Skip to main content

Image generation and editing

Select the right model for text-to-image and image editing.

Text-to-image

We recommend qwen-image-3.0-pro, which excels at generating complex layouts such as newspapers, storyboards, menus, and exam papers, accurately rendering small text down to 10 pixels, supporting multilingual fonts and photographic-level detail. It also supports negative prompts to exclude unwanted elements, can output up to 6 variant images per call, and supports resolutions up to 2048×2048.

When to use wan2.7-image-pro

  • You need resolution above 2K — up to 4096×4096
  • You need precise brand color control
  • You need up to 12 character-consistent images per call in sequential mode

When to use z-image-turbo

  • For image generation only (no editing functionality).
  • Speed or cost is a priority: 10x faster generation at about one-fifth the cost.
  • For photorealistic portraits and product photos.

Image editing

Start with qwen-image-3.0-pro — it supports negative prompts, up to 6 outputs per call, and uses the same model ID for generation and editing. It accepts one to three input images, so you can fuse multiple references. For detailed usage, see Image editing.

When to usewan2.7-image-pro

  • You need more references — up to 9 input images
  • You need bounding-box interactive editing
  • You need character-consistent multi-image sets

Model ID

Use cases

Text-to-image

Editing

Max outputs

Max resolution

qwen-image-3.0-pro

Complex layout generation, small text rendering, multilingual font support

Supported

Supported

6

2048x2048

qwen-image-3.0

A faster version of qwen-image-3.0-pro

Supported

Supported

6

2048x2048

wan2.7-image-pro

Text rendering, brand colors, character-consistent multi-image generation, multi-image editing

Supported

Supported

4 (12 consecutive)

4096x4096 (text-to-image) / 2048x2048 (editing)

wan2.7-image

Same features as the pro version, with faster generation and a lower maximum resolution (2048x2048).

Supported

Supported

4 (12 consecutive)

2048x2048

z-image-turbo

Fast generation, low cost, photorealistic portraits

Supported

Not supported

1

2048x2048

All models

Qwen Image

Model ID

Text-to-image

Editing

Max outputs

Max resolution

qwen-image-3.0-pro

Supported

Supported

6

2048x2048

qwen-image-3.0

Supported

Supported

6

2048x2048

qwen-image-2.0-pro

Supported

Supported

6

2048x2048

qwen-image-2.0-pro-2026-06-22

Supported

Supported

6

2048x2048

qwen-image-2.0-pro-2026-04-22

Supported

Supported

6

2048x2048

qwen-image-2.0-pro-2026-03-03

Supported

Supported

6

2048x2048

qwen-image-2.0

Supported

Supported

6

2048x2048

qwen-image-2.0-2026-03-03

Supported

Supported

6

2048x2048

Wan

Model IDText-to-imageEditingMax outputsMax resolution
wan2.7-image-proSupportedSupported4 (12 consecutive)4096x4096 (text-to-image) / 2048x2048 (editing)
wan2.7-imageSupportedSupported4 (12 consecutive)2048x2048
wan2.6-t2iSupportedSupported41440x1440
wan2.6-imageSupportedSupported41440x1440
wan2.5-t2i-previewSupportedNot supported41440x1440
wan2.5-i2i-previewNot supportedSupported41280x1280
wan2.2-t2i-plusSupportedNot supported41440x1440
wan2.2-t2i-flashSupportedNot supported41440x1440
wan2.1-t2i-plusSupportedNot supported41440x1440
wan2.1-t2i-turboSupportedNot supported41440x1440
Legacy
wanx2.1-imageedit
Available only in the China (Beijing) region
Not supportedSupported11024x1024

Z-Image

Model ID

Text-to-image

Editing

Max outputs

Max resolution

z-image-turbo

Supported

Not supported

1

2048x2048

Token Plan
Model Playground
Fine-tuning
Model Compression
Statistics and Monitoring
Support