Qwen introduces Image 3.0 with longer prompts and finer text rendering

Qwen has introduced Qwen-Image-3.0, a new image-generation model designed for information-heavy visuals, small text and realistic details. QwenTeam announces the model in an official Qwen blog post, describing its central goal as making generated images more useful in practical work.

The company says Qwen-Image-3.0 accepts instructions of up to 4,500 tokens. That capacity is intended to support complex layouts that would be difficult to describe in a short prompt, including newspapers, educational materials, storyboards and multi-panel infographics.

Qwen highlights examples that combine several unrelated subjects in one image, such as a nine-panel grid containing lessons, charts, medical illustrations and business graphics. It also shows nested interface concepts, where a coding environment contains a chat interface, which contains a messaging app and then a poster.

Focus on text and production tasks

Small-text rendering is a major part of the announcement. Qwen says the model can produce text at sizes as small as 10 pixels and handle dense material such as formulas, annotations and newspaper layouts. The company presents examples of academic-style pages with mathematical notation, though it does not provide independent benchmark results or error rates.

Qwen also says the model improves the depiction of textures, including skin, hair, paper and ink-wash paintings. Its editing examples include adding handwritten notes to a book page and restoring damaged sections of traditional artwork.

The model supports native rendering in 12 languages, according to Qwen, and is trained to reproduce familiar interface styles such as websites, games and livestream pages. The company also says it can use internet-connected retrieval for current information, such as a local weather graphic.

Qwen positions the release for workflows where visual output must carry usable information, rather than simply serve as illustration. Potential applications cited by the company include design, education, e-commerce, content production and complex interface mock-ups.