Google has unveiled Lyria 3.5, its latest AI music generation model, set to enhance musicality, lyrics, and vocal quality within Flow Music. This update focuses on increased creative control, allowing users to fine-tune tempo, vocals, drums, and bass, alongside adjustable track lengths ranging from 30 seconds to three minutes.

Lyria 3.5 builds upon its predecessor, Lyria 3, which launched in February and offered song generation up to 30 seconds based on text prompts or photos. The new iteration promises more complex and natural-sounding music, with improved prompt adherence and structural awareness. Google emphasizes that users can still describe the desired mood or theme for their AI-generated songs.

The rollout of Lyria 3.5 in Flow Music began on July 29. Previous versions of Lyria have been integrated into other Google products, including YouTube Shorts for Dream Track and Google Vids for soundtracks, demonstrating Google's ongoing commitment to advancing AI in creative applications.