Google Lyria 3 Pro stretches AI songs to 3 minutes, hits Gemini and enterprise tools

Google launched Lyria 3 Pro on March 25, 2026, a premium tier of the Lyria 3 music-generation model that can now create songs up to three minutes long instead of the previous 30-second limit. The Pro version is rolling out across Gemini, Google AI Studio, and enterprise APIs the same day, according to Google DeepMind’s official blog and TechCrunch AI.

Tracks generated by Lyria 3 Pro are exactly six times longer than the standard Lyria 3 clips—jumping from 30 seconds to a full three minutes. As The Verge AI notes, this leap turns brief snippets into what can now qualify as actual song-length compositions, giving creators room for verses, choruses, and full arrangements instead of mere loops.

Pro access is live in three places: the Gemini mobile and web apps for consumers, Google AI Studio for developers in paid preview, and enterprise endpoints via the Gemini API. Mashable and 9to5Google confirm that free Gemini users can start generating 30-second tracks immediately, while the three-minute Pro output requires a paid tier.

Beyond length, Google says Lyria 3 Pro offers “better creative control” by letting users prompt for specific musical elements—key changes, tempo shifts, instrument choices, and sectional structures. Ars Technica AI frames this as Google sanding down the earlier model’s rough edges, giving musicians enough granularity to guide verses, bridges, and outros instead of accepting whatever the model spits out.

While OpenAI’s Jukebox and Meta’s MusicGen have offered longer clips before, none ship natively inside a mainstream chatbot at Gemini’s scale. By bolting three-minute generation onto an app with hundreds of millions of users, Google leapfrogs rivals on distribution, even if the underlying tech isn’t revolutionary. The move also pressures startups like Suno and Udio that built their pitch around longer-form creation.

Google AI Studio is now offering Lyria 3 Pro endpoints in paid preview, signaling that Google wants third-party apps—from social video editors to marketing platforms—to embed full-length AI soundtracks. If uptake is strong, expect pricing tiers tied to length, audio quality, or commercial licensing, similar to how Google already meters image generation in Vertex AI.

No source detailed new safeguards, but the longer output raises fresh questions about training-data provenance. Rightsholders have already sued AI music startups for allegedly ingesting copyrighted recordings. By extending clips to three minutes, Google increases the odds that generated tracks might closely resemble existing songs, so litigation risk scales with length. Watch for takedown mechanisms and indemnity policies in the coming weeks.

If you need quick, customizable background music for ads, shorts, or prototypes, Lyria 3 Pro is now the fastest path inside Google’s ecosystem. Three minutes is long enough for most commercial use cases, and direct Gemini integration means no extra subscriptions. Serious artists will still prefer human composition or specialized tools, but for content farms and indie creators, Google just turned AI music from a toy into a practical utility.