Stability AI expands creative tools with Audio 3.0
Stability Audio 3.0 marks a step forward in AI music generation by producing longer, cohesive songs and offering a compact variant that runs locally. The flagship model can generate extended, six-minute tracks, giving composers and producers more musical structure to work with than typical short-loop generators.
The release also includes a small on-device model able to generate two-minute tracks. That compact model is built for portability and speed, letting users prototype ideas, sketch full song sections, or create background music without needing an internet connection or cloud resources.
These dual capabilities open up practical benefits: creators get high-capacity generation when they need it and a private, offline option for quick experiments and demos. For independent musicians, podcasters, and small studios, Audio 3.0 can accelerate songwriting, lower production friction, and broaden who can access music technology.
Stability AI's approach — pairing long-form generation with on-device convenience — is likely to spur integrations with music tools and inspire new apps that center creativity, privacy, and accessibility. For many creators, Audio 3.0 represents a meaningful productivity win and a new playground for musical experimentation.