Google announced Wednesday that it will release its music generation model, Lyria 3 Pro, one month after the release of Lyria 3. The Lyria 3 model offered tracks that were 30 seconds long, but the new model allows you to create tracks that are up to 3 minutes long.
The company said the Lyria 3 Pro model offers greater creative control and customization in addition to allowing users to create longer tracks. This model understands track structure better than previous models, so users can also use prompts to specify different elements of a song, such as intros, verses, choruses, bridges, etc.
Google previously introduced music generation functionality in its Gemini app with the Lyria 3 release. The Pro model is also rolled out in the Gemini app, but only paid subscribers can access it. The company is also deploying Lyria 3 Pro in its video editing app Google Vids and ProducerAI, a GenAI-powered music production tool that Google acquired last month.
Additionally, Google is adding music generation capabilities to its enterprise tools with Vertex AI (in public preview), Gemini API, and AI Studio via the Lyria 3 Pro model.
Google emphasized that it used data from partners and permissioned data from YouTube and Google to train this model. He also said the model was not an imitation of the artist. However, if a user specifies an artist in the prompt, the company said it will need “broad inspiration” from that artist to generate the track.
All tracks created using Lyria 3 and Lyria 3 Pro are marked with a SynthID to indicate that AI was used to create this track.
Earlier this week, Spotify released a new tool that allows artists to review songs released under their own name to prevent AI slop creators from misattributing music. Meanwhile, Deezer announced a tool that will allow any streaming service to identify AI-generated music.
