Google has updated its Gemini app to include the Lyria 3.5 model, a tool for generating music directly within the interface. The update allows users to select genres, choose between vocals or instrumentals, and request tracks of varying lengths. New templates assist with background scores or personalised birthday songs. The company states Lyria 3.5 offers more expressive vocals and richer arrangements than previous versions. Access extends beyond the main app to Google Flow Music, Google AI Studio, and Google Vids. Google emphasises that the model trains only on licensed content, distinguishing it from competitors like Suno. Specific details regarding the training dataset remain undisclosed by the firm.
This integration matters because it places a generative audio tool inside a widely used chat interface rather than a niche studio application. Most users access music creation through dedicated platforms, so embedding this capability in Gemini expands the potential audience for AI-generated sound. The licensing claim addresses a significant concern for professional creators who require clear rights to commercialise their work. By offering an API alongside the app, Google also targets developers building their own music applications. The move tests whether general-purpose assistants can effectively handle complex creative tasks without requiring separate software.
- Training data consists solely of licensed material
- Available via API for developer integration
- Templates included for background music and greetings
