Google’s Gemini app now boasts the ability to generate original music, thanks to the integration of Lyria 3, DeepMind’s latest generative AI model. The recent feature allows users to create 30-second musical tracks simply by providing a text description or uploading a photo or video.
Users can now prompt the Gemini app with requests like “a comical R&B ballad about a sock finding its match” and, within seconds, receive a complete track including lyrics, vocals, and production, according to Google. The company highlights three key improvements over previous models: Lyria 3 eliminates the require for users to supply lyrics, generating them automatically based on the prompt. This development underscores the increasing sophistication of AI in creative fields.
With Lyria 3, users also gain “greater creative control over style, voice, and tempo,” and can create “more realistic and musically complex compositions.” When utilizing videos or images, Gemini composes music that “aligns with the visual environment.”
The generated songs, complete with custom cover art created by Nano Banana – presented as singles from an album – are 30 seconds in length and can be downloaded immediately upon creation, along with a link. Google clarifies that the intention isn’t to produce musical masterpieces, but rather to “offer a new form of creative expression, quick and accessible.”
All tracks generated within the Gemini app are embedded with SynthID, Google’s imperceptible watermark for identifying AI-generated content. The feature is currently available in beta for Gemini app users aged 18 and over, in multiple languages, with higher usage limits for subscribers. The launch of Lyria 3 signals a growing trend of AI tools empowering broader creative expression.
