Google has introduced a more advanced version of Lyria 3, Lyria 3 Pro, just a month after the Lyria 3 model.
The new model significantly increases the 30-second limit of the previous version to 3 minutes. This allows users to create full-length music tracks. Lyria 3 Pro better understands musical structure and provides more detailed construction of sections such as intro, chorus and bridge with text commands.
The model is available to paid users in the Gemini app. In addition, the technology is also integrated into tools such as Google Vids and Google's newly acquired ProducerAI. Enterprise users can use the system through Vertex AI, Gemini API and AI Studio.
Regarding copyright and ethical issues, Google says that the model was trained only on authorized data from partner publishers, as well as YouTube and other platforms. The company notes that the system does not directly imitate specific performers, but can take general inspiration from their style when a specific artist's name is specified in user queries.
All music created is tagged with SynthID technology to ensure transparency. Against this backdrop, Spotify has also introduced new tools for monitoring content created with artificial intelligence.
