OpenAI has introduced GPT-Live-1, an innovative ChatGPT version that can listen and speak simultaneously by leveraging full-duplex architecture, according to The Decoder. This advancement enables more natural, real-time conversations without the usual pause between speaking and listening.
The Decoder also reported that complex queries are seamlessly routed to GPT-5.5 in the background, significantly enhancing the quality of responses. GPT-Live-1 is currently available for paying ChatGPT users, while a smaller version is accessible to free account holders. API access is expected to be launched soon.
For Japanese markets, where real-time communication and AI integration are rapidly evolving, GPT-Live-1’s capabilities could accelerate adoption in sectors like finance and customer service, enhancing multilingual and interactive AI solutions.