A text-to-image model using deep learning called Stable Diffusion was released in 2022. Although it can be used for various tasks including inpainting, outpainting, and creating image-to-image translations directed by text prompts, its primary usage is to generate detailed visuals conditioned on text descriptions. In cooperation with a variety of academic researchers and nonprofit groups, the start-up Stability AI created it.

A latent diffusion model, or deep generative neural network, is what is known as stable diffusion. It can function on the majority of consumer hardware outfitted with a modest GPU and at least 8 GB VRAM, and its code and model weights have also been made freely available[4]. This was a change from earlier proprietary text-to-image models like DALL-E and Midjourney, which could only be accessed via cloud services

The start-up company Stability AI provided funding and input into the creation of Stable Diffusion.[The CompVis group at Ludwig Maximilian University of Munich issued the technical licence for the model.[6] Patrick Esser and Robin Rombach, who were among the researchers that created the latent diffusion model architecture employed by Stable Diffusion previously, served as the project’s leaders.[3] As backers of the project, Stability AI also acknowledged Eleuther AI and LAION (a German NGO that put together the dataset used to train Stable Diffusion).[3]

In an investment round led by Lightspeed Venture Partners and Coatue Management in October 2022, Stability AI raised US$101 million.services,