Stability AI Introduces Stable Audio 2.5, the First Audio Model Built for Enterprise Sound Production at Scale

Stability AI launched Stable Audio 2.5, an enterprise-focused audio model for faster music generation and stronger creative control.

The model improves musical structure, prompt adherence, and audio inpainting, and can generate up to three-minute tracks.

Stability AI says it was trained on fully licensed data and can be fine-tuned on a brand’s own sound library.

The company also partnered with amp, part of Landor Group and WPP, to build sound identity tools for enterprise clients.