Stability AI is a generative AI company that trains image, video, 3D and audio models. Its best-known family is Stable Diffusion, text-to-image models that turn a written prompt into a picture. A diffusion model starts from random noise and removes the noise step by step until an image that matches the prompt appears.
Stable Diffusion 3.5 comes as Large (8 billion parameters, the highest quality), Large Turbo (a faster variant that needs only four steps) and Medium (designed to run on consumer hardware). The models can be used in three ways: downloaded and run on the customer's own servers under a license, called through the Stability AI API, which also offers upscaling and image-editing services, or used in Stable Assistant, a web application. Other model families are Stable Audio for music and sound effects, Stable Video, and 3D models that build objects from single images or videos.
The self-hosted Community license is free for organizations with less than US$1 million in annual revenue; larger companies need an Enterprise license. The API is pay per use with prepaid credits.