Amazon introduces Nova Sonic, the latest addition to its Nova family of models. This new model integrates speech recognition, understanding, and generation into a single architecture, enabling AI to more accurately interpret human speech and generate more natural conversational responses.

Amazon states that Nova Sonic outperforms competitor voice models in linking customer commands with various APIs. It can intelligently determine when to retrieve real-time web data, parse proprietary data sources, or select appropriate tools for external applications, offering a cost-effective solution at roughly 20% of the price of OpenAI’s GPT-4.

Nova Sonic is already partially integrated into Alexa+, an enhanced version of Amazon’s voice assistant. A new API within the Amazon Bedrock platform allows developers to build more useful and engaging voice-powered applications.