Qwen-Image Open-Sourced: The Best SOTA Image Generation Model
Qwen-Image: Open-source 20B SOTA image model with superior text rendering & multi-style generation. Outperforms Flux & GPT-Image in benchmarks.
"AI Disruption" Publication 7300 Subscriptions 20% Discount Offer Link.
Now, text generation in images has evolved to this level, and it's open source.
The Tongyi model family has just open-sourced again, this time it's Qwen-Image — a 20 billion parameter image generation model using MMDiT architecture.
This is also the first image generation foundation model in the Qwen series.
Looking at the images generated by Qwen-Image, you can see that one of its main capabilities is complex text rendering.
Like this image of a bookstore bestseller shelf, which contains complex mixed text and image layouts, the accuracy and appropriateness of the text, and even the variations formed by the angles at which books are placed, are flawless.
Generating posters is also effortless.
Beyond text processing, Qwen-Image supports multiple artistic styles in general image generation. From photorealistic scenes to impressionist paintings, from anime styles to minimalist designs, it has mastered them all.
