Introducing Play 3.0 mini – A lightweight, reliable and cost-efficient Multilingual Text-to-Speech model

Play 3.0 mini is a new lightweight, reliable, and cost-efficient multilingual text-to-speech model that can converse in over 30 languages. It achieves a mean latency of 189 milliseconds, making it the fastest model yet, and supports text-in and audio-out streaming via HTTP REST API, websockets API, or SDKs. The model also exhibits significant improvements in audio quality, reliability, and naturalness of speech. Additionally, it features state-of-the-art voice cloning capabilities and is offered at reduced pricing for different business tiers.