Stability AI

Stability AI builds multimodal generative models for images, video and audio and sells access to them as hosted services. The image line is Stable Diffusion 3.5 in Large, Turbo and Medium variants; the video line covers Stable Video 4D 2.0 and Stable Virtual Camera, which turns 2D images into multi-view 3D video; the audio line runs from Stable Audio 2.5 for enterprise sound production to Stable Audio 3.0, described by the company as trained on fully licensed data and released with open weights. The models are reached through an API, through cloud partners, or through applications such as Stable Assistant, Stable Artisan and Brand Studio.

Published deployment options are a self-hosted licence for running models in a customer's own environment, the Stability AI API and cloud marketplaces: Stable Image Services were added to Amazon Bedrock, and Stable Diffusion 3.5 ships as an NVIDIA NIM microservice. A Community Licence covers researchers, developers and businesses under $1M in annual revenue, with an enterprise membership above that. Announced partnerships include Electronic Arts for game development tools and Universal Music Group and Warner Music Group for music creation, and the company states it holds SOC 2 Type II and SOC 3 compliance.

Stability AI Ltd is incorporated in England and Wales with a registered address in London, and Stability AI US Services Corporation in Los Angeles contracts with customers outside the UK, EU and Switzerland. Prem Akkaraju is named as CEO, and the news room reports a closed Series B round that brought total funding to $232 million. Its hosted products are proprietary and run by the company, while the Community Licence governs use of the published weights.