Stability AI has launched Stable Audio 2.5, a generative audio model built specifically for enterprise use. The company claims a new technique allows the model to produce high-quality audio in just eight computational steps, down from 50. Sean Michael Kerner reports for VentureBeat that this breakthrough can cut production time from weeks to minutes. According to Zach Evans, Head of Audio Research at Stability AI, the model reflects a shift toward “enterprise-grade capabilities”. It enables rapid iteration and offers advanced controls for commercial use. New features include audio inpainting, which allows editing of existing tracks, and the ability for companies to train custom models on their own data to align with their brand’s sound identity.