Google introduces Lyria 3, a free AI music generator for Gemini

Google just keeps on releasing new AI tools.

On Wednesday, Google rolled out a new AI music generator called Lyria 3. It's a fairly big upgrade over earlier versions of the model, as it makes music generation a lot easier for users. Lyria 3 can create high-fidelity, 30-second audio tracks in a variety of genres, from short jingles and lo-fi beats to more complex arrangements and songs.

This Tweet is currently unavailable. It might be loading or has been removed.

How to try Lyria 3 for free

Google says that Lyria 3 is available now in the Gemini desktop app (it will be rolling out into the mobile app in the coming days), and it's free for all users 18 and up. Lyria can be used in English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese, with support for additional languages to come.

To get started, simply head to the Gemini app. In the "Tools" dropdown menu, look for the "Music" option, which is accompanied by a music note icon. Google also released a Lyria 3 prompt guide to help users get started.

How Lyria 3 works

Lyria 3 can generate a short 30-second song based on a simple text, image, audio, or video prompt. Users can control the music style, vocals, and tempo all within the prompt. In addition, users no longer have to provide their own lyrics. Lyria 3 can create them based on the prompt. Nano Banana will even generate album artwork to go with it.

If users want to create music to go along with a specific photo or video, they can upload it along with their prompt. For instance, social creators could upload a video and generate a custom music track to go along with it in seconds.

Furthermore, there is a template option with dynamic suggestions for people who'd like to create an AI song but need some inspiration.

According to Google, all tracks that are generated within the Gemini app have a SynthID watermark embedded, which marks content as AI-generated.

While free users will have access to Lyria 3, paying Google AI Plus and AI Ultra subscribers will receive higher usage limits.

What about copyright concerns?

Many musicians and music publishers have criticized the AI industry for training models like Lyria 3 on copyright-protected works without permission or payment, and multiple lawsuits are working their way through the court system based on these concerns. Most AI companies argue that training models on copyrighted works is protected by fair use.

While Big Tech companies generally decline to comment on the specific training materials used for their frontier models, Google did say that it was "mindful" of musical copyrights. A Google blog post announcing the new model said the company has "been very mindful of copyright and partner agreements as we've trained Lyria 3.

Further, Google says that Lyria is not designed to mimic existing artists, and musicians can report content if they believe it violates their rights.

The blog post states:

Music generation with Lyria 3 is designed for original expression, not for mimicking existing artists. If your prompt names a specific artist, Gemini will take this as broad creative inspiration and create a track that shares a similar style or mood. We also have filters in place to check outputs against existing content. We recognize that our approach might not be foolproof, so you can report content that may violate your rights or the rights of others. Additionally, in order to use our products, users must adhere to our Terms of Service and Gen AI prohibited use policies, which prohibit violations of others’ intellectual property and privacy rights.

Disclosure: Ziff Davis, Mashable’s parent company, in April 2025 filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.