Product launch and capabilities
Google announced Lyria 3 Pro, an advanced generative music model designed to produce audio tracks as long as three minutes with built-in awareness of common song structures. The model understands discrete composition elements - including intros, verses, choruses and bridges - enabling users to prompt for specific structures and more complex transitions within a single generated piece.
Availability and developer access
Lyria 3 Pro is being made available in public preview on Vertex AI to support businesses that require on-demand audio generation at scale. Developers can also access the model via Google AI Studio and the Gemini API, providing multiple integration pathways for applications that need programmatic music generation.
Integrations across Google products
The company said it will integrate the extended-generation capability into Google Vids, its AI-driven video creation app, first targeting Google Workspace customers and subscribers to Google AI Pro and Ultra starting this week. In addition, longer-generation functionality from Lyria 3 Pro will be available within the Gemini app for paid subscribers.
Google has also integrated the model into ProducerAI, a collaborative music creation tool aimed at artists, producers and songwriters at various experience levels. ProducerAI is being offered globally to both free and paid users.
Development pathway and collaborators
The firm developed Lyria 3 Pro through its Music AI Sandbox, a set of experimental tools intended for musicians, producers and songwriters. The announcement named Grammy-winning producer Yung Spielburg as a user of Lyria during the composition and production process for the score of the Google DeepMind short film "Dear Upstairs Neighbors." Separately, DJ and producer François K is collaborating with Google using Lyria to create a song that the company said will be released soon.
Training data, output behavior and provenance
Google described the model’s training regimen as relying on materials from YouTube and other sources that Google has a right to use under its terms of service, partner agreements and applicable law. The company said Lyria 3 does not mimic named artists when prompted with creator names, instead taking broader inspiration from its training material.
All outputs from Lyria 3 and Lyria 3 Pro are embedded with SynthID, Google’s watermarking technology for identifying AI-generated content.
Context and implications
The rollout spans enterprise developer tools, consumer creator apps and collaborative production software, reflecting a multi-channel distribution approach. By offering public preview access on Vertex AI and API access through AI Studio and Gemini, Google is positioning Lyria 3 Pro as both an enterprise-scale audio generator and a creative tool for individual creators.