Lyria

Lyria 2 by Google DeepMind is an advanced AI music generation model built to produce high-fidelity, production-ready audio across a wide range of styles. It can generate complete musical pieces, including coherent melody, harmony, rhythm, and long-form structure, making it suitable for full tracks, cues, and underscore. The model supports complex arrangements and instrumentation, covering genres such as orchestral, cinematic, electronic, pop, and experimental music.

A core capability of Lyria 2 is its ability to output professional-grade stems and layered compositions. Users can separate instruments or sections for detailed mixing, sound design, and post-production within standard audio workflows. Lyria 2 responds to text-based prompts that describe mood, genre, tempo, instrumentation, and stylistic requirements, and it can be guided to follow specific musical characteristics such as dynamic contours, motif development, or particular rhythmic patterns.