Skip to main content
Images

Image generation models

Choose a model for text-to-image generation, image editing, and more.

Text-to-image

Start with qwen-image-3.0-pro — it excels at complex layout generation (newspapers, storyboards, menus, exam papers), 10-pixel fine-grained text rendering, multilingual fonts, and photographic detail. It supports negative prompts to exclude unwanted elements and up to 6 variants per call. Up to 2048×2048. Switch to qwen-image-3.0 for the same capabilities with faster generation.

When to use wan2.7-image-pro instead

  • 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 instead

  • You only need generation (no editing)
  • Speed or cost is the priority — 10x faster, ~1/5 the price
  • Realistic portraits and product shots

Not sure which fits?

Try models in Try AI — same prompt, compare results. Style preference is personal.

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.

When to use wan2.7-image-pro instead

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

Erasing and segmentation

Remove people, objects, text, or watermarks from an image and fill in the erased area → image-erase-completion (China North 2 (Beijing) only, free trial). Generate a pixel-level mask for each person → image-instance-segmentation (China North 2 (Beijing) only, free trial). See Image erase completion and Person instance segmentation.
Model IDUse this whent2iEditMax outputMax resolution
qwen-image-3.0-proComplex layouts, fine-grained text rendering, multilingual fonts, photographic details, negative prompts✓✓62048×2048
qwen-image-3.0Faster variant of qwen-image-3.0-pro✓✓62048×2048
wan2.7-image-pro4K output, brand colors, multi-image sets, multi-image editing✓✓4 (12 sequential)4096×4096 (t2i) / 2048×2048 (edit)
wan2.7-imageSame capabilities, faster generation, up to 2K✓✓4 (12 sequential)2048×2048
z-image-turboFast generation, low cost, realistic portraits✓12048×2048

All models

In the console, the model list, search results, versions, prices, and rate limits follow the current region. Models differ by region, and some are available only in specific regions, so you can only compare models within the same region. Before you choose a model, check the region's tab in Model prices. See Regions and endpoints.
The model pages in the console, including model details and model comparison, show the models, prices, and rate limits of the region selected in the upper-right corner. A model offered in one region may not be listed in another. If you switch regions while a model detail page or the comparison list is open, a model that is not available in the new region is flagged, and the page shows the other regions where the model is available. The comparison page stays open even if only one model remains.
Model IDt2iEditMax outputMax resolution
qwen-image-3.0-pro✓✓62048×2048
qwen-image-3.0✓✓62048×2048
qwen-image-2.1-pro✓✓62048×2048
qwen-image-2.1-turbo✓✓62048×2048
qwen-image-2.0-pro✓✓62048×2048
qwen-image-2.0-pro-2026-06-22✓✓62048×2048
qwen-image-2.0-pro-2026-04-22✓✓62048×2048
qwen-image-2.0-pro-2026-03-03✓✓62048×2048
qwen-image-2.0✓✓62048×2048
qwen-image-2.0-2026-03-03✓✓62048×2048
qwen-image-2.1 series supports up to 10 reference images for editing, and can output images with an alpha channel based on the prompt.
Model IDt2iEditMax outputMax resolution
wan2.7-image-pro✓✓4 (12 sequential)4096×4096 (t2i) / 2048×2048 (edit)
wan2.7-image✓✓4 (12 sequential)2048×2048
Model IDt2iEditMax outputMax resolution
z-image-turbo✓12048×2048
Previous generation models. We recommend Qwen Image 3.0 or Wan 2.7 for new projects.

Wan

Model IDt2iEditMax outputMax resolution
wan2.6-t2i✓4~1440×1440
wan2.6-image✓*✓4~1440×1440
wan2.5-t2i-preview✓4~1440×1440
wan2.5-i2i-preview✓41280×1280
wan2.2-t2i-plus✓4~1440×1440
wan2.2-t2i-flash✓4~1440×1440
wan2.1-t2i-plus✓4~1440×1440
wan2.1-t2i-turbo✓4~1440×1440
* Requires enable_interleave=true and stream=true. See text-to-image guide.
Wan 2.1 model IDs depend on the region. In Singapore (ap-southeast-1), use wan2.1-t2i-plus and wan2.1-t2i-turbo. In China North 2 (Beijing) (cn-beijing), use wanx2.1-t2i-plus and wanx2.1-t2i-turbo.
Model IDt2iEditMax outputMax resolution
vidu/vidu-image_reference2image✓✓14K tier
Available in Singapore only. See Vidu.

Learn more