Generate images from descriptions or edit with one to three references. Qwen Image 3.0 Pro offers 1K and 2K tiers, eight aspect ratios, prompt rewriting and PNG or JPEG output.
Pro takes the same inputs as Qwen Image 3.0, a description or one to three reference images, and exposes the same tiers, shapes, output formats and prompt controls in our workspace.
We do not claim Pro produces better results than the standard model. What differs is the quote: Pro is quoted above the standard model, and its 2K tier is quoted above its 1K tier, where the standard model quotes both tiers the same. Run the same prompt through both and keep the output you prefer.
Set the tier, shape, output format and the optional negative prompt and seed exactly as on the standard model. Prompt rewriting is on by default, and the quote updates with the tier, the reference count and the number of images you request.
Two tiers, eight shapes, one to three references, PNG or JPEG output, prompt rewriting, a negative prompt and an optional seed.
The 2K tier is quoted above 1K on this model, which makes the tier a deliberate cost decision instead of a silent one.
Upload one to three of your own images and describe the change, with every reference quoted separately from the first.
Try a chosen prompt and references on Pro, then compare the result with your draft before publishing.
Run one prompt through both family products at the same tier and judge the difference on your own material.
Combine up to three of your own images into one render when the job needs a specific subject, product or layout.
Use a button on this page. It opens the right workspace with this model already selected, in the language you are reading now.
Write what should happen in the shot or image. Concrete subjects, framing and action give the model more to work with than short labels.
Choose length or quality tier, shape and sound where offered, then read the credit quote for those exact settings before you generate.
Generation uses account credits. Cost depends on the model, quality and output size or duration. Reference images, video or audio may add charges in some modes. Check the workspace quote for your selected inputs and settings before generating; the pricing page lists credit packages.
See credit packagesThe standard model has a lower quote at both tiers. You can start there and try Pro with the same prompt and settings to compare results.
Results depend on your prompt and references. Compare the outputs for your task; a higher quote does not guarantee a better result.
Per image, by tier, with each reference image added on from the first. The exact amount for your settings is shown before you confirm.
Between one and three per editing task, and each one is quoted separately.
Upload JPEG, PNG or WebP, up to 10 MB per image; finished files are PNG or JPEG.
Other models on this platform that do a similar job, or that pair well with this one.
Qwen Image 3.0 renders an image from a description or edits from one to three of your own references, at the 1K or 2K tier, with PNG or JPEG output.
Read the guideNano Banana 2.1 covers both workflows on one model: images from a description, or edits driven by up to ten of your own reference images, at 1K, 2K or 4K.
Read the guideThe second selectable GPT Image 2.5 product: the same tiers, shapes, background control and credit prices as Flare, with this guide focused on editing your own images.
Read the guide