Google released Lyria 3 Pro on March 25, 2026, extending its AI music model's maximum output from 30 seconds to three minutes — a six-fold increase. The model is now available in the Gemini app for paid subscribers, Google Vids, and ProducerAI, a music production platform Google recently acquired.

Why it matters

For working musicians and content creators, 30-second AI clips are demos, not deliverables. A three-minute generation limit crosses the threshold into usable song territory — enough for a full pop structure with intro, verse, chorus, and outro. Lyria 3 Pro shifts the use case from experimentation to actual production workflows, which is why the rollout targets professionals via ProducerAI and the Gemini paid tier rather than free users.

Track length and structural control

Lyria 3 Pro lets users specify discrete song elements — intros, verses, choruses, and bridges — rather than prompting for a vague style. Google DeepMind built this structural awareness into the model so outputs reflect intended composition rather than ambient loops. All Lyria 3 Pro audio is watermarked and designed to avoid mimicking existing artists, addressing the rights concerns that have shadowed AI music tools since 2023.

Competitive context

The launch comes roughly a month after the original Lyria 3 release, signaling an accelerating release cadence. Google is competing directly with Suno, Udio, and ElevenLabs music tools, all of which already support full-length song generation. By embedding Lyria 3 Pro into Vertex AI and the Gemini API as well, Google is positioning the model as an enterprise music layer, not just a consumer feature — a framing none of its direct competitors have matched at scale.