🤖 OpenAI releases the GPT-Live speech model, which supports full-duplex real-time dialogue. OpenAI releases the new generation speech model GPT-Live, which uses the full-duplex architecture to realize real-time dialogue while speaking and listening. It will be launched to ChatGPT users around the world from now on.

🤖 OpenAI releases the GPT-Live speech model, which supports full-duplex real-time conversations. OpenAI releases a new generation speech model, GPT-Live, which uses a full-duplex architecture to achieve real-time conversations while speaking and listening. It will be launched to ChatGPT users around the world from now on. The model can process input and output synchronously, respond naturally in conversations, allow users to interrupt or pause at any time, and call GPT-5.5 in the background to complete complex tasks such as search and deep reasoning. GPT-Live is divided into two versions: GPT-Live-1 and GPT-Live-1 mini, which will become the default model of ChatGPT Voice for paid and free users respectively. The new version also adds…