Lyria 2

Overview

Freshness note: Model capabilities, limits, and pricing can change quickly. This profile is a point-in-time snapshot last verified on February 15, 2026.

Lyria 2 is Google’s music generation model line for AI-assisted composition and soundtrack experimentation. Google’s current Gemini API and pricing docs still frame the Lyria family as paid-tier developer-facing music generation capability, even though access remains more limited and creator-oriented than mainstream text models.

Capabilities

The model is designed for music-centric generation tasks, including composition-style prompting, soundtrack ideation, and creative iteration. It is useful in prototyping, early content scoring, and experimental creative exploration pipelines.

Technical Details

As an audio generation model, token contextWindow and maxOutput are represented as 0 in this repository and should be treated as N/A. Real evaluation should focus on musical coherence, timbre, and structure quality under target use conditions.

Pricing & Access

Google’s current pricing docs list Lyria 2 at $0.06 per second of generated audio, with access on paid developer tiers where the model is enabled. Availability and workflow controls remain more constrained than on mainstream Gemini text models, so teams should confirm channel and region support before planning production dependency.

Best Use Cases

Best for soundtrack ideation, interactive music experiments, and early-stage creative development where fast iteration matters more than full DAW-level control.

Comparisons

Compared with Suno and Udio tool ecosystems, Lyria 2 is better viewed as a model capability layer rather than a complete creator platform. Practical model choice should be based on output quality tests against your target genre and workflow constraints.