Google Launches Lyria 3 Pro: AI Music Generation Up to Three Minutes

Google launched Lyria 3 Pro on March 25, 2026, a major upgrade to its AI music model. Where its predecessor Lyria 3 generated 30-second tracks, the new model can create full songs of up to three minutes with structured sections like intros, verses, choruses, and bridges.

Lyria 3 Pro is available in the Gemini app for paid subscribers (Google AI Pro and Ultra) and is integrated across Vertex AI, Google AI Studio, Gemini API, ProducerAI, and Google Vids. The model generates vocals, lyrics, and cover art in a single operation, with all outputs watermarked with SynthID for traceability. Multilingual music is supported.

The jump from 30 seconds to three minutes is technically demanding. A musical idea with a beginning, middle, and end requires a fundamentally different level of coherence and structure than a short clip. The fact that Lyria 3 Pro delivers this with creative control over song structure is a clear signal that Google's music AI is now competitive with dedicated services like Suno and Udio.

For businesses, the integration with Google Vids is particularly interesting. Videos with tailored AI music can be produced without royalty costs or licensing risk, which matters for everything from internal training to marketing campaigns.

OpenAI shut down Sora two days earlier. The timing is unlikely to be coincidental: Google is filling a void in the creative AI suite just as a competitor's video ambitions stall.