Latest models
- Stable Audio 3.0 (May 20, 2026) Current
A family trained on licensed data. Small, Small SFX and Medium are open weights on Hugging Face; Large is available through the API. Medium reaches about six minutes per track. Source: Stability AI ↗
Checked October 7, 2026. See the full AI model tracker.
What we like
- Trained on fully licensed data
- Small SFX, Small and Medium are open weights on Hugging Face
- Variable-length generation; Medium reaches about 6 minutes 20 seconds
Watch out for
- The Large model needs the API or enterprise self-hosting
- More technical to set up than a simple web app
- Check the Community or Enterprise licence for your use
Stable Audio is Stability AI’s family of audio generation models. Stable Audio 3.0 was released on May 20, 2026.
What is current
Stability says the 3.0 family is trained on fully licensed data. Stable Audio 3.0 Small SFX, Small and Medium are open weights, and Small and Medium are on Hugging Face. Large is available through the Stability AI API and enterprise self-hosting. Small can generate up to two minutes; Medium reaches up to 6 minutes 20 seconds. Stability says you own your outputs and can distribute and commercialise them under the Stability AI Community License, or the Enterprise License for organisations with more than $1 million in revenue. See the model tracker.
Who should use it
Producers and developers who want open weights, licensed training data and the option to run locally, especially for instrumentals and sound effects. For vocal songs from a prompt see Suno.