Google launches Lyria 3 Pro music generation model
The new model creates tracks 6x longer than its predecessor and offers granular control over song sections.
Google has released Lyria 3 Pro, a significant upgrade to its music generation model just one month after Lyria 3 debuted. The headline feature is a massive increase in output length, jumping from 30-second snippets to full, 3-minute compositions. Beyond duration, the Pro model offers vastly improved creative control, allowing users to specify structural elements like intros, verses, choruses, and bridges within their prompts. This represents a move from generating generic audio loops to creating structured, dynamic musical pieces.
Access to Lyria 3 Pro is gated behind a paid Gemini subscription, marking a clear monetization strategy. The model is also being integrated across Google's ecosystem, including the Google Vids video editor and the newly acquired ProducerAI music production tool. For enterprise users, it's available in public preview through Vertex AI, the Gemini API, and AI Studio. Google trained the model on data from partners and permissible YouTube/Google data, emphasizing it doesn't mimic specific artists but can take "broad inspiration" from named styles. All generated tracks are watermarked with SynthID for transparency, arriving alongside industry efforts from Spotify and Deezer to label AI-generated content.
- Generates 3-minute tracks, a 6x increase from the previous 30-second limit
- Allows granular prompt control over song structure (intros, verses, choruses, bridges)
- Rolling out to paid Gemini tiers, Google Vids, ProducerAI, and enterprise Vertex AI
Why It Matters
Moves AI music from short demos to usable compositions, giving creators and businesses a powerful new audio tool.