Lyria 3 Is Here: Google DeepMind Powers AI Music Inside Gemini App
Lyria 3, Google DeepMind’s newest artificial intelligence (AI) music generating model, was unveiled on Wednesday. According to the tech giant located in Mountain View, the new model is incorporated into the Gemini app for iOS and Android and is intended to assist users in creating high-quality music songs by utilizing multimodal generative AI and straightforward language cues. Lyria 3 is marketed as a tool for creators that makes music production easier for digital storytelling, social media, and video.
Features of Lyria 3
According to a blog post by Google DeepMind, Lyria 3 drives new music composition features within the Gemini app. By specifying the preferred genre, mood, speed, or instrumentation, users can create creative tunes that last up to 30 seconds. Text prompts are converted into professionally produced audio clips with layered instrumentation and a polished sound output by the music creation model.
Introducing Lyria 3, our new music generation model in Gemini that lets you turn any idea, photo, or video into a high-fidelity track with custom lyrics.
From funny jingles to lo-fi beats, you can create custom 30-second soundtracks for any moment.
See how it works. ?
— Google Gemini (@GeminiApp) February 18, 2026
Lyria 3 is said to provide a number of improved features over its predecessors. First of all, it can produce lyrics automatically in response to the prompt, giving you greater control over aspects like tempo, style, and singers. In addition, it provides stronger coherence over longer pieces, richer instrumentation, and enhanced audio quality.
In addition to generating words to tracks, Lyria 3 in Gemini enables users to compose music based on images and videos. An photograph of a sunset or a brief journey video, for instance, can be utilized as creative input, and Lyria 3 will produce a soundtrack that complements the mood and ambiance of the image. Custom AI-generated cover art can be included with each generated track, which facilitates the packaging and distribution of artists’ work.
Additionally, Google is expanding Lyria’s capabilities to YouTube Dream Track, allowing users to test out AI-generated music for short videos and other types of content. Dream Track uses comparable generative AI technologies to allow composers to experiment with various musical styles and themes.
In response to concerns regarding the legitimacy of AI-generated media, Google has included its SynthID watermarking technology into Lyria 3 outputs. For transparency and ethical use, it subtly adds metadata to generated audio to assist identify it as AI-generated content.
The model, according to the corporation, is a component of their larger drive toward multimodal AI, in which the creation of text, images, music, and video all coexist on a single platform.
According to the tech behemoth, Lyria 3 is currently accessible on desktop computers in English, German, Spanish, French, Hindi, Japanese, Korean, and Portuguese through the Gemini app for all users over the age of 18. Over the next days, it will be made available via the Gemini app for iOS and Android. Higher restrictions are available to Google AI Plus, Pro, and Ultra members.
