Stability AI Stable Audio 3.0 music generation
Key Questions
What is Stable Audio 3.0 and what can it generate?
Stable Audio 3.0 is a family of AI music models from Stability AI capable of generating up to six-minute songs. It includes open-weights variants, an on-device small model, and an audio effects variant.
What data was used to train Stable Audio 3.0?
The models were trained on fully licensed data to support commercial and enterprise use cases. This builds on prior Stable Audio releases with improved quality and capabilities.
Are there open-weight versions of Stable Audio 3.0 available?
Yes, Stability AI is releasing open-weights variants along with a smaller on-device model and enterprise licensing options. Community builds are also encouraged.
How does Stable Audio 3.0 improve on previous versions?
It extends earlier releases by enabling longer song generation up to six minutes and adding specialized audio effects. The family focuses on licensed data for safer, higher-quality outputs.
What licensing options exist for Stable Audio 3.0?
Enterprise licensing is available alongside open-weight releases, allowing both individual creators and businesses to use the models. This supports a range of community and commercial applications.
Stable Audio 3.0 family for 6-minute song generation, open-weights variants + on-device small model, licensed data, enterprise licensing, community builds, audio effects variant. Builds on prior Stable Audio releases with improved capabilities.