OpenAI: New Voice Models for Natural Live AI Conversations

OpenAI has launched new voice models, GPT-Live-1 and GPT-Live-1 mini, designed for more natural live conversations. These models aim to make AI interactions feel much closer to talking with another person. The significant upgrade is their full-duplex capability, meaning they can listen and speak simultaneously. This allows users to interrupt, ask follow-up questions, or correct the AI without waiting for it to finish. This also enables features like instant live translation. GPT-Live-1 mini will become the new default in ChatGPT, replacing the existing Advanced Voice Mode. The more powerful GPT-Live-1 will be available for paid subscribers. These new models integrate speech-to-text, response generation, and text-to-speech more tightly. They connect directly to OpenAI’s latest text models, including GPT-5.5. This helps the AI understand context better and pause naturally. What's interesting is their ability to maintain context over much longer discussions, with internal tests showing conversations lasting 30 to 40 minutes. The new interface can also present information visually, like charts or lists, alongside the conversation. The bottom line is that these advancements aim to make interacting with AI voice assistants far more seamless and intuitive for everyone.