Lyria 3.5 Music Generation Model Now Available on Gemini App and API

Google has expanded its music-generation model Lyria 3.5 to the Gemini app and API, making it easier for users to create high-quality music with actual song structure. The model was first released in July through Google Flow Music, but now it's available to a wider audience.

Lyrira 3.5 allows users to specify a song's temporal structure directly within the input, including vocals, timed lyrics, and full instrumental arrangements. Users can also pass up to 10 images alongside text to build music from colors, landscapes, or the mood of a subject. The model uses an approach that applies latent diffusion to a time-directional latent representation of audio.

However, it's essential to note that Google advises checking whether generated tracks match the intended outcome. Additionally, Lyria 3.5 does not support interactive editing, where a generated clip is gradually refined through additional prompts. The model generates full-length songs with intros, verses, choruses, and bridges.