Google’s Gemini AI assistant can now generate original music tracks up to 30 seconds long, a feature powered by the company’s latest Lyria 3 AI model, Google announced Wednesday. The capability allows users to create music from text prompts, uploaded images, or videos.
The new feature, currently available on the desktop version of Gemini and rolling out to the mobile app in the coming days, enables users to generate both lyrical songs and purely instrumental pieces. Google stated that users over the age of 18 will have access to the music generation tool in multiple languages. Alongside the audio, Gemini will also generate a custom cover art using Google’s Nano Banana image creation model, allowing users to share visually complete tracks.
The rollout of Lyria 3 within Gemini comes as Google seeks to demonstrate the commercial potential of its AI investments and compete with OpenAI’s ChatGPT. Google’s Gemini 3 model, launched in November, was widely praised by investors and users, prompting OpenAI CEO Sam Altman to reportedly initiate a “code red” to accelerate improvements to ChatGPT, according to reports.
Google is implementing usage limits for the free version of the music generator, allowing users to create 10 tracks per day. Paid subscribers will have higher limits, ranging from 20 to 100 tracks daily depending on their subscription tier. Google asserts that users will retain ownership of the music they generate and that the system incorporates filters to prevent copyright or privacy violations.
The introduction of AI-powered music creation tools is occurring alongside similar developments from other tech companies. Apple this week announced Playlist Playground, a feature for Apple Music that uses AI to create playlists from text prompts, complete with cover art and descriptions. This feature, included in iOS 26.4 beta, is slated for wider release this spring and directly competes with a similar offering from Spotify.
News of Google’s Lyria 3 integration into Gemini prompted a brief dip in Spotify’s stock price, whereas shares of Sirius XM Holdings also declined, according to Bloomberg Intelligence analysts, who do not anticipate a significant disruption to Spotify’s business but suggest the move may spur Spotify to launch its own AI-powered mixing feature.
The music industry has shown caution regarding generative AI, with Universal Music Group, Warner Music Group, and Sony Music Entertainment filing lawsuits against AI startups Suno AI and Uncharted Labs (Udio AI) in 2024 alleging copyright infringement. Warner Music has since reached an agreement with Suno, and both Warner and Universal have established licensing agreements with Udio to allow their applications to continue operating.
Google stated that Lyria 3 is trained on music that YouTube and Google have the rights to use, adhering to their terms of service, partner agreements, and applicable laws. The company also claims to have safeguards in place to prevent the AI from replicating the style of specific artists, instead using artist names as “broad creative inspiration” to generate music with a similar style or mood.
