Stability AI releases new audio models, creates six-minute songs

Stability AI has unveiled its latest audio generation models, Stability Audio 3.0, capable of producing professional-grade music up to six minutes and twenty seconds long. This represents a significant leap from its previous iterations, doubling the generation length and offering improved musical structure and melodic coherence. The company is making smaller models available with open weights for broader use, while larger, more advanced models will be accessible via API and paid services. Stability AI is also focusing on professional music offerings, hiring industry veterans and leveraging its partnerships with major music labels, ensuring its models are trained on fully licensed data.