What is Google DeepMind Lyria 3.5?

Lyria 3.5 is Google DeepMind’s flagship generative audio model designed for musical composition. It synthesizes full-length 44.1 kHz stereo audio tracks with expressive vocals, complex instrumental arrangements, and defined musical sections directly from text or image inputs.

Where can creators and developers access Lyria 3.5?

Lyria 3.5 is deployed in three primary surfaces: the consumer Gemini app (web and mobile), Google AI Studio for interactive developer prototyping, and the Gemini API for enterprise and software application integration.

How does Lyria 3.5 differ from Suno v4 and Udio 1.5?

Lyria 3.5 features tighter prompt adherence on structural timestamps, native multimodal visual prompting (accepting up to 10 images), hardware-integrated SynthID audio watermarking, and lower API pricing ($0.08 for a full 3-minute track) compared to commercial credit packages.

Can you iteratively edit stems or specific sections in Lyria 3.5?

No. Lyria 3.5 operates as a single-turn generative model. While you can dictate tempo, instruments, and timestamps in your prompt, you cannot currently in-paint or replace individual vocal stems without generating a new track from a seed.

How does SynthID audio watermarking protect Lyria 3.5 generations?

SynthID embeds an imperceptible, pseudorandom frequency signature into the audio spectrogram before waveform decoding. The mark persists through lossy compression (such as 128 kbps MP3 encoding), analog re-recording, and modest pitch adjustments, allowing automated verification without degrading listener fidelity.

Google Lyria 3.5 Deep Report: Benchmarks, Gemini API Guide & Prompts

A technical analysis of Google DeepMind’s Lyria 3.5, deployed across Gemini, Google AI Studio, and the Gemini API. Covers 44.1 kHz stereo synthesis, 3-minute full-song structural coherence, SynthID audio watermarking, multimodal image-to-audio prompting, and benchmarks against Suno v4 and Udio 1.5.

# Google Launches Lyria 3.5: Multimodal Music Creation in Gemini, API Benchmarks, and Prompt Architecture Google DeepMind launched **Lyria 3.5**, expanding its audio generation foundation model across the consumer **Gemini app**, **Google AI Studio**, and the **Gemini API**. Following its initial summer deployment inside Google Flow Music, this release opens commercial-grade generative music to developers and creators globally. Lyria 3.5 generates full-length songs up to three minutes long in **44.1 kHz stereo audio**. It introduces multimodal visual conditioning, structural timestamp controls, multilingual vocal synthesis, and integrated **SynthID** audio watermarking. This report evaluates Lyria 3.5 through empirical benchmarks against Suno v4 and Udio 1.5, details its underlying latent diffusion architecture, tests its prompt grammar, and provides complete Gemini API integration code. --- ## Architectural Profile: Latent Audio Diffusion and SynthID Music generation requires m...