Stability AI offers Stable Diffusion, a family of AI models that generate images from text descriptions, available as open-source weights and via a commercial API. The latest version, Stable Diffusion 3.5, accepts text prompts and produces high-resolution images across styles including photography, painting, and 3D rendering.

Stable Diffusion is a latent diffusion model, first released in August 2022, that converts text prompts into detailed images using a deep learning process. The model family has expanded through multiple versions including SDXL, Stable Diffusion 3, and Stable Diffusion 3.5 Large (a 3.5 billion parameter model), each improving prompt adherence, image quality, and text rendering. Models are available for self-hosting, via Stability AI's Developer Platform API, and through third-party cloud providers including Amazon Bedrock.