Cartesia has launched Sonic-3.6, a new text-to-speech model designed to deliver more natural AI-generated speech across 44 languages.
The model builds on improvements informed by teams using Cartesia’s technology, with a focus on more lifelike voice interactions while maintaining the speed needed for real-time applications, according to the company.
Sonic-3.6 is available in beta. Cartesia said it ranks first on Artificial Analysis’ provider and controlled voice streaming leaderboards. Its previous model, Sonic-3.5, and Sonic-3.6 also hold the top two positions on the controlled benchmark.
The launch comes three months after Sonic-3.5, with the company saying the latest release incorporates feedback from teams building on its platform. Cartesia develops AI products including Sonic and Ink.
Source: LinkedIn
