Google DeepMind Launches Lyria 3 Pro Music Generator
- •Google DeepMind debuts Lyria 3 Pro for high-fidelity music generation up to three minutes long
- •Model features structural awareness allowing users to prompt for specific elements like verses and choruses
- •Integration spans Vertex AI, Gemini, and Google Vids for professional and consumer use cases
Google DeepMind has officially launched Lyria 3 Pro, an evolution of its custom music generation model designed to provide artists and developers with greater structural control over high-fidelity audio. Unlike previous iterations limited to shorter clips, this Pro version supports tracks up to three minutes in length while maintaining a sophisticated grasp of musical composition. Users can now direct the AI to generate specific song components—such as intros, verses, and choruses—facilitating more complex transitions and professional-grade arrangements.
The rollout strategy emphasizes broad accessibility across the Google ecosystem. For enterprise users, Lyria 3 Pro enters public preview on Vertex AI, enabling the scaling of bespoke soundtracks for gaming and marketing. Developers can access the model via Google AI Studio, while everyday creators can utilize its capabilities within Google Vids and the Gemini app. This widespread integration suggests Google is positioning Lyria as a foundational layer for creative audio production across both professional and consumer-grade tools.
Addressing industry concerns regarding intellectual property, Google emphasizes that Lyria 3 Pro was trained using authorized materials and includes safeguards against direct artist mimicry. Every output is embedded with SynthID, an imperceptible watermark that identifies the content as AI-generated. By collaborating with musicians like François K and Yung Spielburg, Google aims to frame generative AI as a collaborative tool rather than a replacement for human creativity, focusing on refinement and precision in the creative workflow.