# Lyria 3: AI model fact sheet

- **Provider:** google
- **Released:** 2026-02-18
- **Capabilities:** music-generation, audio-generation

Google DeepMind's most advanced music generation model, launched in the Gemini app. First Lyria model with full vocal generation, automatic lyric creation, and multimodal inputs (text, image, video). Produces 30-second tracks with distinct verses, choruses, and bridges. Supports 8 languages for vocal generation. No public API yet, available through the Gemini app and YouTube Dream Track.

Source: real side-by-side outputs, pricing and specs at https://rival.tips/models/lyria-3