Google Gemini has expanded its AI capabilities to include music generation, leveraging the new Lyria 3 audio model. Users can now create 30-second songs from simple text prompts, or by providing images or videos to set the mood.
The Lyria 3 model, developed by Google DeepMind, allows for creative musical compositions based on user descriptions. Examples include generating a "comical R&B slow jam about a sock finding their match" or a "sea shanty about the dangers of AI slop."
Users can also choose to remix existing musical presets or have the AI compose a tune to match the mood of supplied visual content. The feature is available to all Gemini users, with paid subscribers of Gemini Plus, Pro, or Ultra potentially accessing higher usage limits, though these are not specified.
Google aims for this feature to be a "fun, unique way to express yourself," rather than for producing professional musical works. All generated tracks will include invisible watermarks powered by SynthID to identify them as AI creations. The Lyria 3 model is also being integrated into YouTube Shorts for creators.
The update is rolling out on the Gemini web app for users aged 18 and above in English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese, with mobile app availability to follow. Google plans future expansions for improved quality and broader coverage.