Google Unleashes Lyria 3.5 Music Model on Gemini App and API

Google has made its Lyria 3.5 music model available to all users of the Gemini app and API, allowing them to generate full songs at 44.1kHz from text or images. The model, which first launched in Google Flow Music on July 29, can produce tracks lasting several minutes, complete with verses, choruses, and bridges. Lyria 3 Clip, a related model, always produces 30-second clips.

The API's documentation notes that developers can supply up to ten images, write their own lyrics using section tags, and use timestamps to specify when instruments enter or exit. However, the model has safety filters in place to block requests for specific artist voices or the generation of copyrighted lyrics.

This move comes as a Munich court ruled against Suno on July 31 over its memorized works and infringing outputs. The court asserted jurisdiction over training done outside the EU and concluded that US fair use did not apply. Google has not confirmed whether this ruling is connected to Lyria's release, but notes that every generated track carries an imperceptible SynthID watermark.