GPT-Live: OpenAI's Real-Time Voice Model for ChatGPT

OpenAI has launched GPT-Live, a new voice model for ChatGPT, allowing for more natural and fluid conversations. This feature is rolling out globally across iOS, Android, and the web, marking a major upgrade to ChatGPT Voice. What's interesting is that GPT-Live supports simultaneous listening and speaking. Users can interrupt ChatGPT mid-sentence or change topics without waiting for the AI to finish its response. The system also uses conversational cues, like brief acknowledgments, to make interactions feel more like a real discussion. The AI can now continue speaking while handling complex tasks in the background, whether it's a web search or processing multiple requests. This reduces awkward pauses and improves conversation flow. Paid subscribers will get GPT-Live-1 as the default voice model, while free-tier users receive GPT-Live-1 mini. The bottom line is this aims to make voice interaction a primary and more human-like way of using AI assistants.