Google announced on Wednesday the release of Lyria 3 Pro, a new music generation model, just a month after the launch of Lyria 3. The updated model enables users to create tracks up to three minutes long, an improvement over the 30-second limit of Lyria 3.
The company highlighted that Lyria 3 Pro allows for greater creative control and customization. Users can specify various musical elements in their prompts, such as intros, verses, choruses, and bridges, benefiting from the model’s improved understanding of track structure compared to its predecessor.
Previously, Google introduced music generation capability to the Gemini app with Lyria 3. The Pro model is also available in the Gemini app, but access is limited to paid subscribers. Additionally, Lyria 3 Pro is being integrated into Google’s Vids video editing app and ProducerAI, the AI-powered music production tool Google acquired last month.
Google is expanding music generation capabilities to enterprise tools using the Lyria 3 Pro model in Vertex AI (currently in public preview), the Gemini API, and AI Studio.
Google clarified that Lyria 3 Pro was trained with data from partners and permissible sources like YouTube and Google, stressing that it doesn’t imitate specific artists. However, it draws “broad inspiration” from specified artists in user prompts to generate tracks.
All tracks created with Lyria 3 and Lyria 3 Pro are marked with SynthID to indicate they were made with AI.
Earlier this week, Spotify introduced tools allowing artists to review songs released under their name to prevent misattribution due to AI. Deezer also launched tools to help streaming services identify AI-generated music.