Lyria 3 Pro launches 3-minute song tools in Gemini API and AI Studio

• Google says launch post puts two music models live today: Lyria 3 Pro for full-song generation and Lyria 3 Clip for 30-second outputs.
• The practical change is structure control: feature thread shows creators can map intros, verses, choruses, and bridges instead of stitching short clips into a longer track.
• Access is split across products, with rollout details saying developers can use the models in Google AI Studio via API while paid subscribers get them inside the Gemini app.

Google's launch post frames this as a two-model release: Lyria 3 Pro for full songs and Lyria 3 Clip for shorter 30-second generations. The launch is available in the Gemini API and in a new music experience inside Google AI Studio, with more product detail in Google's overview.

The big workflow upgrade is length plus arrangement. Google DeepMind's feature thread says Lyria 3 Pro can generate tracks up to three minutes long, and the demo arrangement demo shows a timeline broken into Intro, Verse 1, Chorus, and Bridge. That means musicians can prompt toward a song form in one pass rather than generating isolated ideas and editing them together afterward. Lyria 3 Clip covers the shorter end of that workflow with 30-second generations, according to model split.