Google Launches Lyria 3 Pro: Full 3-Minute AI Songs with Verse/Chorus Structure
Google DeepMind has launched Lyria 3 Pro, its most capable AI music generation model to date — and it marks a meaningful leap beyond anything the company has shipped in the audio space before. Where previous models could produce short clips or loops, Lyria 3 Pro generates complete, structured songs up to three minutes long, with coherent verse and chorus arrangements built from nothing more than a text prompt. The model is rolling out this week to Gemini app paid subscribers, Google Workspace customers, and Google AI Pro and Ultra subscribers.
The pace of iteration here is striking. Google went from Lyria 2 to Lyria 3 to Lyria 3 Pro in under a year, and each release has expanded what's possible. With Lyria 3 Pro, the focus shifts from experimentation to production: the model integrates directly with Google Vids and enterprise Workspace products, which means content teams and businesses can now generate broadcast-quality music as part of their existing workflows rather than through a separate creative tool.
For developers and content professionals building on Google's stack, Lyria 3 Pro signals a new category of media automation. Long-form, structurally coherent AI music in a production environment changes the calculus for everything from marketing video soundtracks to app experiences — and positions Google's creative AI offerings as a serious counterpart to its model and coding tools.