Stability AI · audio
Stable Audio 3
Instrumental tracks and soundscapes cleared for commercial use, from a few seconds to about six minutes.
Text to music
About Stable Audio 3
Stable Audio 3 is Stability AI's music and sound model, priced at a flat $0.10 per generation. It runs as text to music; fal describes the Medium model as generating stereo music from text prompts.
Unlike the music models priced by length, it charges one price per track. A typical generation takes about 1 min.
Good for
- Instrumental tracks and soundscapes in stereo
- Long beds of music at a price that does not grow with length
Getting good results
- Type the length in seconds; leave it empty and the provider uses its default. The price is the same either way.
- Use the negative prompt, under Advanced, for qualities to avoid, and pick WAV or FLAC there when you need a lossless file.
Stable Audio 3 questions
Does a longer track cost more?
No. Every track costs $0.10, whatever length you type.
What is it trained on?
fal notes that Stable Audio 3 Medium was trained on fully licensed data.