Stability AI, the company behind the text-to-image AI model Stable Diffusion, has just released an early preview of Stable Diffusion 3.
“With Stable Diffusion 3, we strive to offer adaptable solutions that enable individuals, developers, and enterprises to unleash their creativity, aligning with our mission to activate humanity’s potential,” Stability AI wrote in a blog post.
According to the company, Stable Diffusion 3 improves performance for multi-subject prompts, image quality, and spelling abilities.
For example, the photo shown here was generated by Stable Diffusion 3, using the prompt: “Epic anime artwork of a wizard atop a mountain at night casting a cosmic spell into the dark sky that says ‘Stable Diffusion 3’ made out of colorful energy.” The blog post also showcased images created by Stable Diffusion that have text written on objects, such as street signs, a bus destination sign, and a chalkboard.
The Stable Diffusion 3 models will range from 800 million to 8 billion parameters. “This approach aims to align with our core values and democratize access, providing users with a variety of options for scalability and quality to best meet their creative needs,” Stability AI wrote.
The company also noted that it introduced several new safeguards in this release in order to improve responsible AI development. No details about the specific safeguards were revealed.
Anyone looking to get started with the early preview must sign up for the waitlist. The company says its primary reason for the early preview is to gain insights from users so that it can improve the performance and safety prior to the actual release.