Stability AI launches Stable Diffusion 3.5 to enhance image generation models

Stability AI has launched Stable Diffusion 3.5, a significant update to its text-to-image generative AI technology. This version aims to improve upon the previous release, which did not meet the company's standards. The new models offer enhanced customization and quality. Stable Diffusion 3.5 includes three variants: Large, Large Turbo, and Medium, each designed for different user needs. The models are available under an open license for non-commercial use and for commercial use by smaller entities. The update features improved prompt adherence and new techniques like Query-Key Normalization for better model training. Stability AI plans to introduce ControlNets for more precise control in professional applications in the future.