AI Tool
Stable Audio 3
Stability AI · Open weights (free) / API
Stability AI's open-weight generative audio model — fast local inference, six-minute generations, and ComfyUI integration. The caveat: no vocals, no lyrics.
7.8
GoodSpecifications
- Type
- Generative audio model
- Output
- Up to 6-minute generations
- Inference
- Seconds on Apple M4; on-device capable
- Caveat
- No vocals or lyrics
Last verified 2026-06-12
FAQ
Can Stable Audio 3 generate vocals?
No — it produces instrumental music, foley, and sound effects, but no vocals or lyrics.
Is Stable Audio 3 free?
The weights are open and can run locally; an API is also available.