ChatGPT's Voice Mode: Interrupt & AI Listens

ChatGPT's new voice mode now lets you interrupt mid-sentence, and the AI will actually listen. This update fundamentally changes how the assistant handles spoken conversation. OpenAI released two new voice models: GPT-Live-1 for paid subscribers and GPT-Live-1 mini for free users. They are available on iOS, Android, and ChatGPT.com. The previous voice mode worked sequentially, but GPT-Live uses full-duplex architecture, meaning the model listens and speaks at the same time. This allows you to cut in, ask it to slow down, or redirect the conversation without waiting for a pause. The assistant also uses short verbal acknowledgements like "got it" to signal it's following along. When a question needs deeper reasoning, the model hands off to GPT-5.5 in the background while keeping the conversation flowing. Web search, real-time translation, file and image uploads, and info cards for weather, stocks, and sports scores are all included. Paid users get GPT-Live-1, while free users receive GPT-Live-1 mini, which has the same conversational architecture but lower capability. Voice latency is now a subscription differentiator. OpenAI has also added safety guardrails, including monitoring for self-harm signals and parental controls to disable voice mode for teen accounts. This development impacts how millions interact with AI every day.