Meet Stable Audio 3.0, the model family built for artistic experimentation with open-weight models
Stability AI released Stable Audio 3.0, a family of open-weights music models trained on licensed data, with variable-length generation up to six minutes.
- Stable Audio 3.0 Small SFX, Small, and Medium are open-weights and available on Hugging Face; Large is via API and self-hosting.
- Small generates up to two minutes, while Medium and Large generate more than six minutes.
- Outputs are owned by users and can be commercialized under the Stability AI Community License, with Enterprise license for organizations over $1M revenue.
- Supports LoRa fine-tuning, audio inpainting, and on-device full music composition for Small.
Musicians and developers can now build and commercialize music generation models with open weights and licensed training data, enabling on-device and customizable audio creation.
Sources
Read this as text
Back to the AI news