Independent benchmarking reports and evaluations

Cartesia AI Launches Sonic, a 135ms State Space Voice Model for Real-Time Intelligence

May 29, 2024, 05:24 PM

Cartesia AI has launched its first model, Sonic, a state-of-the-art voice model designed to deliver real-time, multimodal intelligence across devices. The Sonic model boasts a blazing fast latency of approximately 135ms, and even as low as 100ms, making it highly efficient for text-to-speech applications. This release marks a significant step in Cartesia AI's mission to develop new, efficient architectures for intelligence, moving away from traditional transformer models. The Sonic model is based on a new state space model architecture, which has been praised for its low latency and high-quality performance. The model is expected to revolutionize the field by providing a viable alternative to existing architectures and enabling real-time intelligence on every device. The Sonic model and its API are now available for public consumption, and users are encouraged to try it out.