Google's new Lyria 3.5 AI model brings better vocals, lyrics
What's the story
Google has introduced its latest AI model for music generation, Lyria 3.5, through its platform Google Flow Music. The model promises advancements in melody complexity and emotional expression. According to Google, it features improved prompt adherence and structural awareness, ensuring lyrics align closely with user inputs. The vocals are described as more realistic and emotionally expressive, with enhanced pronunciation for a richer listening experience.
User customization
Users can now adjust tempo and duration of tracks
Lyria 3.5 offers users expanded creative options to personalize their music. Tracks can now range from 30 seconds to three minutes, with adjustable tempo and duration settings.
Google emphasized that users can modify specific elements such as vocals, drums, and bass to suit their preferences.
The model allows for songs based on detailed prompts describing themes or vibes, including requests like "a goofy R&B jam about a lonely sock."
Model evolution
Lyria 3 laid the groundwork for the latest model
Lyria 3.5 expands on the foundation set by its predecessor, Lyria 3, which launched in February 2026.
The earlier model allowed users to generate short songs using text prompts or photos.
Google integrated Lyria 3 into platforms like Dream Track for YouTube Shorts and Google Vids, enabling video creators to craft custom soundtracks.
The latest update continues Google's push for advancing AI-driven music creation tools globally.