Google launches Lyria 3.5 AI music model with richer vocals

Google has launched Lyria 3.5, the latest version of its artificial intelligence music generation model, promising more natural-sounding tracks and more expressive, emotionally nuanced vocals. The model began rolling out on July 29 within the company's AI song-creation platform, Google Flow Music.

Developed by Google DeepMind, the new model delivers what the company describes as significant advancements in musicality, lyrics, and vocal quality. Improvements include richer melodic structures, higher-quality lyrics with better adherence to user prompts, and more precise control over a track's tempo and duration. A new feature called "Selective Section Painting" allows users to edit specific parts of a generated track, rewrite lyrics, or change instruments without altering the entire song. Tracks can now run from 30 seconds to three minutes in length.

The launch continues the rapid evolution of the Lyria model line, which Google first unveiled in November 2023. An earlier version, Lyria 3, launched in February 2026 within the Gemini app, allowing users to generate 30-second songs from text or image prompts. A more advanced Lyria 3 Pro model followed in March, enabling the creation...