Artificial intelligence startup Stability AI Ltd. has announced the third generation of Stable Diffusion.
Announcing Stable Diffusion 3, our most capable text-to-image model, utilizing a diffusion transformer architecture for greatly improved performance in multi-subject prompts, image quality, and spelling abilities.
Today, we are opening the waitlist for early preview. This phase… pic.twitter.com/FRn4ofC57s
— Stability AI (@StabilityAI) February 22, 2024
The new AI model will feature numerous enhancements, including improved image quality and additional capabilities for prompts. Developers have particularly focused on its ability to generate text on images without errors.
Stable Diffusion 3 will be available in several versions with varying numbers of customizable parameters, ranging from 800 million to 8 billion.
Heavy models capable of creating vivid and complex scenes will require significant computational power. Lighter versions can be run on a computer without cloud services.
“This approach is aimed at aligning with our core values and democratizing access. We provide users with various options to best meet their creative needs”, stated representatives of Stability AI.
Open source code will allow direct access to the models’ underlying architecture. Those interested in testing Stable Diffusion 3 can apply to the waitlist.
In August 2023, Stability AI launched a free alternative to ChatGPT.