Google DeepMind Launches Lyria 3 AI Music Generator Inside the Gemini App

Google DeepMind released Lyria 3 on February 18, 2026, making it the first major AI music generation model to be embedded directly in a consumer AI assistant. The model is available in beta through the Gemini app to users aged 18 and older, generating 30-second music tracks with full lyrics from text prompts.

Capabilities

Lyria 3 generates complete audio tracks — melody, harmony, rhythm, instrumentation, and vocals with lyrics — from natural language descriptions. Users can describe a desired genre, mood, instrumentation, and lyrical theme, and the model produces a 30-second track. The system also accepts image and video uploads as creative input, analyzing visual content to generate complementary music. Alongside the audio track, Lyria 3 generates AI album art, providing a complete audio-visual package from a single prompt.

Google describes Lyria 3 as its "most advanced" AI music generation model. The integration into the Gemini app, rather than a standalone tool, positions music generation as a natural extension of multimodal AI assistance — sitting alongside text generation, image analysis, and code assistance in the same interface.

SynthID Watermarking

All Lyria 3 outputs are watermarked with SynthID, Google's imperceptible AI-content identification technology. SynthID embeds identifying information directly into the audio waveform in a way that is inaudible to human listeners but detectable by automated verification tools. The watermark survives common audio transformations including format conversion, compression, and moderate editing, providing a persistent chain of provenance from generation to distribution.

The SynthID integration reflects a broader industry movement toward tracking AI-generated content at the point of creation rather than relying on after-the-fact detection. As generative AI models become capable of producing increasingly realistic audio, the ability to distinguish AI-generated music from human-created recordings has significant implications for copyright enforcement, platform moderation, and consumer trust.

YouTube Dream Track Integration

Beyond the Gemini app, Google is integrating Lyria 3 into YouTube Dream Track, a tool for YouTube Shorts creators that generates background music and soundtrack elements directly within the video creation workflow. Dream Track integration gives creators access to AI-generated music without leaving the YouTube platform, potentially reducing reliance on stock music libraries and simplifying the production pipeline for short-form video content.

Industry Implications

Lyria 3's launch into a consumer product used by hundreds of millions of people raises immediate questions about the music industry's response. Previous AI music tools — including Google's own earlier Lyria versions — have operated in more limited contexts. Placing a capable music generator inside the Gemini app makes AI music creation as accessible as asking a chatbot a question. The music industry's relationship with AI-generated content remains contested, with major labels pursuing legal action against some AI music companies while cautiously engaging with others. Google's existing licensing relationships with music publishers through YouTube may provide a framework for navigating these tensions.