Google announced on Wednesday that it is releasing Lyria 3 Pro, a new music generation model, just one month after the launch of Lyria 3. The new model allows users to create tracks up to three minutes long, a significant increase from the 30-second tracks offered by the previous Lyria 3 model.
The company stated that, in addition to enabling longer tracks, the Lyria 3 Pro model provides better creative control and customization. Users can specify different structural elements of a musical piece directly in their prompt, such as intros, verses, choruses, and bridges, as the model has an improved understanding of track structure compared to its predecessor.
Google first introduced music generation to the Gemini app with the Lyria 3 release. The new Pro model is also rolling out in the Gemini app, but access will be limited to paid subscribers. Google is also integrating Lyria 3 Pro into its Google Vids video editing app and ProducerAI, a generative AI-powered music production tool that Google acquired last month.
Furthermore, Google is adding music generation capability to its enterprise tools, including Vertex AI in public preview, the Gemini API, and AI Studio, through the Lyria 3 Pro model.
Google emphasized that it trained this model using data from its partners and permissible data from YouTube and Google. The company also clarified that the model does not mimic specific artists. However, if a user names an artist in a prompt, the model will take broad inspiration from that artist’s style to generate a track.
All tracks created using either Lyria 3 or Lyria 3 Pro are marked with SynthID, a watermark denoting that AI was used to create the music.
This news follows other recent developments in the industry regarding AI-generated music. Earlier this week, Spotify released new tools that let artists review songs released under their name to prevent misattribution by AI content creators. Meanwhile, Deezer has launched tools to help any streaming service identify AI-generated music.