qwen-image-3.0-pro

Qwen-Image-3.0-Pro was officially released in August 2026 and is the flagship version of the Qwen-Image series, positioned as an “image model for complex visual content production.” Its core upgrades revolve around three directions: “rich content, realistic details, and solid knowledge.” It supports ultra-long instruction input of up to 4.5K tokens and can organize multi-level complex layouts within a single image, such as newspapers, exam papers, storyboards, and menus. It supports rendering of approximately 10px small text, formulas, and realistic materials, approaching photographic quality. It natively supports 12 languages, more than 20 fonts, and interface simulations for webpages, games, livestreaming, and more. The model supports both text-to-image and image editing; for image editing, it can accept 1 to 3 reference images and generate up to 6 images in a single run.

Context

0 tokens

Max output

0 tokens

Input

llm image

Output

image