{"slug": "stable-audio-3", "title": "Stable Audio 3", "summary": "Stable Audio 3 is a family of fast latent diffusion models (small, medium, large) designed for variable-length audio generation and editing, capable of producing several minutes of sound. The models use a novel semantic-acoustic autoencoder for efficient generation and support inpainting for targeted audio editing, with post-training to improve quality and speed. Trained on licensed and Creative Commons data, the small and medium models can run on consumer hardware, generating audio in under two seconds on an H200 GPU.", "body_md": "Computer Science > Sound\n[Submitted on 18 May 2026]\nTitle:Stable Audio 3\nView 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 sounds. We also support inpainting, enabling targeted audio editing and the continuation of short recordings. Our latent diffusion models operate on top of a novel semantic-acoustic autoencoder that projects audio into a compact latent space, enabling efficient diffusion-based generation while preserving audio fidelity and encouraging semantic structure in the latent. Finally, we run adversarial post-training to both accelerate inference and improve generation quality, reducing the number of inference steps while improving fidelity and prompt adherence. Stable Audio 3 models are trained on licensed and Creative Commons data to generate music and sounds in less than a 2s on an H200 GPU and less than a few seconds on a MacBook Pro M4. We release the weights of small and medium, that can run on consumer-grade hardware, together with their training and inference pipeline.\nReferences & Citations\nLoading...\nBibliographic and Citation Tools\nBibliographic Explorer (What is the Explorer?)\nConnected Papers (What is Connected Papers?)\nLitmaps (What is Litmaps?)\nscite Smart Citations (What are Smart Citations?)\nCode, Data and Media Associated with this Article\nalphaXiv (What is alphaXiv?)\nCatalyzeX Code Finder for Papers (What is CatalyzeX?)\nDagsHub (What is DagsHub?)\nGotit.pub (What is GotitPub?)\nHugging Face (What is Huggingface?)\nScienceCast (What is ScienceCast?)\nDemos\nRecommenders and Search Tools\nInfluence Flower (What are Influence Flowers?)\nCORE Recommender (What is CORE?)\narXivLabs: experimental projects with community collaborators\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\nBoth individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.\nHave an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.", "url": "https://wpnews.pro/news/stable-audio-3", "canonical_source": "https://arxiv.org/abs/2605.17991", "published_at": "2026-05-20 15:10:05+00:00", "updated_at": "2026-05-20 16:37:43.751376+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "open-source", "research", "products"], "entities": ["Stable Audio 3", "H200 GPU", "MacBook Pro M4", "Creative Commons"], "alternates": {"html": "https://wpnews.pro/news/stable-audio-3", "markdown": "https://wpnews.pro/news/stable-audio-3.md", "text": "https://wpnews.pro/news/stable-audio-3.txt", "jsonld": "https://wpnews.pro/news/stable-audio-3.jsonld"}}