Alibaba Qwen releases Qwen-Image-3.0, a third-gen image model built around realism

通义千问发布Qwen-Image-3.0：主打"真实"的第三代图像生成模型

Qwen-Image-3.0 puts realism front and center. It accepts prompts up to 4.5k tokens and can generate complex layouts—newspapers, exam sheets, 3×3 infographics—in one shot. Text renders down to 10px, and it handles full LaTeX paper pages plus lifelike skin textures. The model supports 12 languages, 100+ art styles, and real-time web retrieval, targeting design and education workflows. The post doesn't disclose parameter count, inference cost, or release format, so I'd hold off until real-world tests land.

Why it matters: Alibaba Qwen ships a new image model with a clear 'realism' positioning, backed by concrete specs: 4.5k-token prompts, 10px text rendering, full LaTeX page output. Not scoring 85+ because we only have the announcement — no third-party comparisons yet, and 'realism' claims need...