Amazon unveils its newest advanced AI voice model, Nova Sonic, for real-time conversations

Amazon has launched Nova Sonic, a new AI voice model designed for real-time conversational applications, competing with Gemini Live, Copilot's Voice or OpenAI's Advanced Voice Mode. Nova Sonic uses a "unified model architecture" to streamline processes like speech recognition, speech-to-text, response generation, and text-to-speech conversion into one approach. This integration enhances conversational AI by accurately detecting users' tone and delivering natural-sounding responses.

Nova Sonic is available through Amazon's Bedrock developer platform, allowing developers to incorporate it into various applications, including customer service bots and AI agents in sectors such as travel, education, and healthcare. Some features of Nova Sonic are already part of Alexa Plus, the recently unveiled new advanced AI voice assistant from Amazon.

Additionally, Amazon introduced Nova Reel 1.1, an updated video model with better quality and reduced latency compared to Nova Reel 1.0. This new version can maintain consistent visual styles across various clips, enabling the creation of longer, coherent videos up to two minutes long.