# Lyria 2

> Audio AI Model by Google — available on Layer

Google DeepMind's high-fidelity 48 kHz music model with fine-grained creative control.

Lyria 2 from Google DeepMind generates high-fidelity 48 kHz stereo music with rich vocals and complex compositions. The model supports text-based music generation with fine-grained control over key, BPM, and stylistic elements, plus the ability to extend existing tracks while maintaining tonal consistency. It spans genres from pop to funk to classical, and generates vocals in multiple languages. All output is watermarked with SynthID for responsible AI identification.

## Specifications

| Property | Value |
|----------|-------|
| Provider | Google |
| Category | Audio |
| Price | 2.8 / generation per generation |

## Capabilities

- Auto Duration

---

[Try Lyria 2](https://next.app.layer.ai/me/session/new?baseModelId=LYRIA_2) | [All Audio Models](https://layer.ai/models/audio) | [All Models](https://layer.ai/models)