Google has released Lyria 3.5, a new version of its artificial intelligence (AI) model for generating music, making it available through the Gemini app and Gemini API.

The company describes Lyria 3.5 as its highest-quality music generation model to date, with improvements aimed at producing more expressive vocals and more detailed musical arrangements.

The model allows users to generate music by selecting or describing a genre and choosing between vocal and instrumental tracks. Google has also introduced templates intended to help users create different types of music, including background tracks and personalised songs for occasions such as birthdays.

Google said the model could be used for applications including background music for videos, brand jingles and personalised ringtones.

Lyria 3.5 is available globally through the web version of Gemini and its mobile app. Google is also making the model available through Flow Music for artists and AI creators, Google AI Studio for developers and technologists, and Google Vids, the company's video creation and editing platform.