Woofun AI reports that the Alibaba Qwen team launched Qwen Image 3.0 on July 21. The model processes up to 4500 tokens, a 4.5x increase over its predecessor, enabling single-run generation of complex layouts like newspapers and infographics. It renders 10px text, fine details such as hair strands, LaTeX formulas, and supports 12 languages natively. API trials are active on chat.qwen.ai without announced pricing, while model weights and technical reports remain undisclosed.