Stability AI, the company behind Stable Diffusion’s text-to-image AI model, has just released an early preview of Stable Diffusion 3.
“With Stable Diffusion 3, we aim to offer customizable solutions that enable individuals, developers and companies to unleash their creativity, in line with our mission to activate humanity’s potential,” wrote Stability AI in blog post.
According to the company, Stable Diffusion 3 improves multi-subject query performance, image quality, and spelling capabilities.
For example, the photo shown here was generated by Stable Diffusion 3, using the query: “Epic anime artwork of a wizard on top of a mountain at night casting a cosmic spell on the dark sky that says ‘Stable Diffusion 3’ made of colorful energy.” The blog post also shows images created by Stable Diffusion with text printed on objects, such as street signs, a bus stop sign and a billboard.
Stable Diffusion 3 models will have a range of 800 million to 8 billion parameters. “This approach aims to align with our core values and democratize access, providing users with a range of scalability and quality options to best meet their creative needs,” Stability AI wrote.
The company also noted that it introduced several new safeguards in this release to improve responsible AI development. No details on specific safeguards have been disclosed.
Anyone who wants to get started with the early review must sign up for the waiting list. The company says the primary reason for the early preview is to get user insight to improve performance and security before the actual release.