Models

Gemini API surface for Lyria 3 Clip, Lyria 3 Pro and Lyria RealTime music generation.

Gaudio Lab source-separation model for isolating vocals and instruments from mixed audio.

Song-generation model presented by the HappyShrimp product for complete-song creation.

API for music catalog ingestion, metadata and AI-assisted music search workflows.

Released component of the HeartMuLa open music-foundation-model stack.

Released component of the HeartMuLa open music-foundation-model stack.

Released component of the HeartMuLa open music-foundation-model stack.

Contrastive audio-text model trained for music retrieval and classification.

Contrastive audio-text model trained for music and speech retrieval and classification.

Google DeepMind music model for generating music and lyrics from prompts and other inputs.

Google's preview model for short music clips and loops.

Google music model for full songs and developer generation workflows.

Google's preview flagship for structured full-length songs.

Experimental Google model for continuously steerable instrumental-music streaming.

Beatoven.ai model for background music and sound-effect generation.

Non-autoregressive MAGNeT checkpoint for 10 seconds of music.