Gemini Now Lets You Create Music With AI for Free.
Google Gemini can help you write text, generate images and videos, and write code. Now, this AI bot can also create music, rivaling services like Suno in generating melodies based on a simple text request. This update is made possible by the new Lyria 3 audio generation model, which is built into Gemini starting today. Developed by Google DeepMind, Lyria has already been available in other Google products (such as Vertex AI and YouTube Shorts) to a limited number of users, but this is the first time Google is making this model widely available to everyone.
Lyria works the same way as creating images or videos: simply describe what you want, and the AI will do the rest. You might want to hear “a fun, slow R&B song about a sock finding its mate” (as in Google’s example), or perhaps “a nautical song about the dangers of low-quality AI materials”—the choice is yours.
When you tap the “Create Music” button in the Gemini app, you also have the option to select an existing track for a remix, rather than starting from scratch—if, for example, you’re short on inspiration. These presets cover everything from folk ballads to Latin pop, so you can see the musical range they offer. You can also provide Lyria 3 with an image or video in Gemini, and the program will create a composition that matches the mood of the content you provide, including music and lyrics. Google gives an example of a situation where you provide Gemini with several photos of your dog, and the program then creates a song about your pet and their adventures.
Currently, each track is limited to 30 seconds, and while music creation is available to all Gemini users, subscriptions to Plus, Pro, or Ultra will have higher usage limits (though the limits aren’t specified). According to the official blog post, the goal “isn’t to create a musical masterpiece, but to provide you with a fun and unique way to express yourself.” You won’t be able to create your own AI-generated group on Spotify , but you can create a few entertaining tracks for your own (or others’) enjoyment.
I haven’t had a chance to test this feature yet, but I listened to a few examples provided by Google. They sound fairly standard and mundane, exactly as you’d expect from the average result of processing a huge amount of training audio data —as if song genres had been reduced to their most common components and repackaged.
Google says all created tracks will contain invisible watermarks generated by SynthID, identifying them as AI creations, and you’ll be able to upload audio tracks to Gemini and verify them with SynthID. The updated Lyria 3 will also be available in the Dream Track music editor for YouTube Shorts creators. Lyria 3 is already being rolled out to Gemini and is available to users over 18 in English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese. It’s available first in the web app, and will be coming to the mobile app in the next few days. Google says it plans to expand its “quality and reach” in the future.