Artem Voloskovets
The model improves vocal expression, musical arrangements, and overall audio quality and can generate complete songs rather than short music clips.
What Lyria 3.5 can generate
Lyria 3.5 supports full-length tracks with structured compositions including:
- verses;
- choruses;
- bridges;
- vocals and instrumental sections.
Audio is generated as 44.1 kHz stereo, making the output suitable for music, video soundtracks, games, ads, and other production workflows.
| Feature | Lyria 3.5 |
| Generation | Full-length songs |
| Structure | Verses, choruses, bridges |
| Vocals | Supported |
| Audio quality | 44.1 kHz stereo |
| Access | Google AI Studio, Gemini API, Gemini app |
For developers
The key update is Gemini API access. Developers can integrate Lyria 3.5 directly into applications instead of relying on the Gemini interface.
This enables products that generate complete music tracks from prompts, including automated soundtrack generation, music creation tools, game audio, and personalized songs.
Compared with short-form music generators, support for multi-section song structures is particularly important: the model can maintain an arrangement across verse → chorus → bridge transitions instead of producing only a single musical fragment.
Lyria 3.5 is now available in Google AI Studio and the Gemini API, while users can access the model directly through the Gemini app.
Artem Voloskovets
