On September 4, 2026, Google announced that its Lyria 3.5 music generation model was available in the Gemini app and Gemini API. The update focuses on creating music tracks based on users’ ideas.
Google claims more expressive vocals and richer arrangements. This is the developer’s assessment. In the app, the company offers a choice of vocal or instrumental tracks, as well as shorter or longer compositions. This gives users several ways to define a track’s format before generating it.
Practical context: When trying out background music, it is useful to start with an instrumental version and compare results of different lengths. When assessing the update, listen to the vocals and arrangement separately: these are the aspects Google highlights in its description. For a specific project, it makes sense to check how well the result matches the desired mood and tempo before including it in the final edit.
Sources
Event date: 2026-09-04. Original source date: 2026-09-04.
