- Google officially announced “Lyria 3.5,” the latest version of its music generation AI model “Lyria.”
- It is the highest-quality music generation AI model to date, delivering more expressive vocals and richer musical arrangements, allowing users to produce tracks with higher fidelity than ever before.
- In addition to genre selection and prompt input, users can now choose whether tracks feature vocals or instrumental-only arrangements. New templates tailored for various purposes, ranging from background music to birthday songs, have also been introduced.
On Friday, September 4, 2026 (local time), Google officially announced “Lyria 3.5,” the latest version of its music generation AI model “Lyria.”
“Lyria 3.5” is the highest-quality music generation AI model to date. It delivers more expressive vocals and richer musical arrangements, enabling users to produce tracks with higher fidelity than ever before.
“Lyria 3.5” will be rolled out through the Gemini app and the Gemini API. In the Gemini app, alongside genre selection and prompt input, users can now choose whether the music should include vocals or be instrumental only. Additionally, new templates tailored for specific purposes—ranging from background music to birthday songs—are now available.
Furthermore, users can now choose whether to generate short or long tracks, making it easier to bring ideas to life, whether for video accompaniment, original brand jingles, or custom ringtones.
In addition to the Gemini app and Gemini API, “Lyria 3.5” will be rolled out to artists and AI creators through the AI video editing service “Google Vids” and the AI music production service “Google Flow Music.” It will also be made available to developers via “Google AI Studio.”
Source: Google (https://blog.google/innovation-and-ai/products/gemini-app/better-tracks-lyria-gemini/)