Stable Audio 3
arxiv.org
May 20, 2026
2 min read
🔥🔥🔥🔥🔥
49/100
Summary
Computer Science > Sound [Submitted on 18 May 2026] Title:Stable Audio 3 View PDF HTML (experimental)Abstract:Stable Audio 3 is a family of fast latent diffusion models (small, medium, large) for variable-length audio generation and editing. Since our models can generate several minutes of audio, variable-length generations are key to avoid the cost of producing full-length generations for short s...