Stability AI releases Stable Audio 3.0 for music up to 6 minutes 20 seconds

Stable Audio 3.0 comes in four variants, spanning 459M parameters for Small and SFX, 1.4B for Medium, and 2.7B for Large. The three smaller variants have open weights on Hugging Face and allow commercial use under the Stability AI Community License. All training audio was licensed. The models support inpainting, variable-length generation, and extension.

Visit original source