Google announces the integration of its most advanced AI music generation model, Lyria 3, into the Gemini App, with global rollout starting February 19. Developed by Google DeepMind, this technology is the latest version in the Lyria series. Unlike previous tools that could only generate pure music, this upgrade enables the creation of complete songs.
Lyria 3 can not only compose music automatically but also generate original lyrics and produce more realistic vocal performances. Users can freely set the genre, mood, tempo, and singing style, even adjust instrument arrangements, making the creative process closer to their mental image. This means that from inspiration to finished product, there’s no need for complex music production workflows.
30-Second High-Quality Songs Starting from a Single Prompt
Using Gemini, users simply access the “Create music” feature, input a description of the desired song, and the system will generate a 30-second high-fidelity music piece. Whether it’s energetic rock, laid-back lo-fi, or humorous advertising-style tunes, all can be created with simple prompts.
Examples showcased by Google include an energetic 90s skate punk-style song complaining about a roommate not doing the dishes, paired with fast drums and male vocals; lo-fi music inspired by a sunset photo; and cinematic background music generated based on video content. These cases demonstrate Lyria 3’s ability to understand context and atmosphere, making music creation intuitive and fun.
Automatic Cover Art and Built-in Watermark for Responsible Creation
In addition to music, Gemini will also generate a dedicated cover image for each song. The cover is created by Google’s image model Nano Banana, giving the作品 a more complete feel when shared. Users can directly download the music and cover for social media posts, short videos, or personal projects.
As AI content becomes more widespread, Google emphasizes the importance of responsibility. Every song generated with Lyria 3 will embed an invisible digital watermark called SynthID, allowing identification of AI-generated content. Users can also upload audio files for detection to verify if the watermark is present, showing Google’s active efforts in ensuring verifiable AI music.
Global Launch with Higher Quotas for Subscribers
Lyria 3 is now available on the desktop version of Gemini and will gradually expand to mobile devices. The feature is open to users worldwide aged 18 and above, supporting eight languages: English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese.
In terms of usage limits, free users can generate about 10 songs per day, while Gemini AI Plus, Pro, and Ultra subscribers enjoy higher generation caps. This design allows casual users to experience the tool easily while providing more capacity for heavy creators.
Community Buzz and YouTube Expansion: AI Music Fully Penetrates
Since its launch, Lyria 3 has quickly sparked discussions on X (formerly Twitter). Many users share meme songs, K-pop style experimental tracks, or create personalized theme songs for daily life. The ability to generate content instantly with low barriers is attracting a large number of content creators and amateur music enthusiasts.
Meanwhile, Google also announced that YouTube’s Dream Track feature is now globally available, enabling creators to use similar Lyria technology to generate AI music clips, further expanding possibilities for video content creation.
Gemini Becomes a Comprehensive AI Creation Platform
With Lyria 3, Gemini is no longer just a text or image tool but a multimedia creation platform encompassing text, images, videos, and music. The line between AI and professional creative software is gradually blurring, significantly lowering the barrier to entry.
When a melody can be generated instantly from just a few words, everyone can have their own original theme song. Google encourages users to keep experimenting and sharing—perhaps an AI-driven personal music era is already beginning.