Google unveiled the AI model Lyria 3 Pro, which generates three‑minute music tracks, but it is paid.

Google releases a new music model — Lyria 3 Pro

*Google announced the launch of Lyria 3 Pro, an advanced version of its AI model for music generation that debuted in July as Lyria 3.*

What’s changed?
Feature Lyria 3 Lyria 3 Pro
Track length up to 30 seconds up to 3 minutes
Creative control basic extended (structure refinement)
Composition settings limited full-featured requests: intros, verses, choruses, bridges, etc.
Lyria 3 Pro “better understands track structure” than its predecessor, enabling the creation of more complex and longer musical works.

Where can it be used?
* Gemini – the model is already available in the app; the Pro version will only be visible to paid subscribers.
* Google Vids – the video editor now includes Lyria 3 Pro as a music‑creation tool.
* ProducerAI – the service acquired by Google last month also received support for the model.
* Enterprise solutions – through Vertex AI (pre‑public version), Gemini API, and AI Studio, music can be generated with Lyria 3 Pro.

How was the model trained?
Google used partner data and licensed YouTube content but claims it “does not imitate specific artists.” If a user specifies an artist’s name in a prompt, the model takes a “broad inspiration” from them rather than copying their style.

All compositions created with either Lyria 3 or Lyria 3 Pro are marked with a SynthID tag, indicating that the piece was generated by artificial intelligence.

Thus, the new version of Lyria opens up more flexible possibilities for creating musical tracks up to 3 minutes long and provides advanced tools for structuring compositions.