Skip to main content

ChatGPT Desktop Now Talks Like a Real Person: GPT-Live Voice Goes Live

OpenAI just gave the ChatGPT desktop app a serious upgrade. The client, which already merged with Codex, now officially introduces ChatGPT Voice mode powered by the GPT-Live model. In the latest version, you can start, check, or adjust tasks using voice across three sections: Chat, Work, and Codex. The idea? Move multi-threaded collaboration from typing to speaking.

This feature has been a long time coming. OpenAI first added speech recognition to ChatGPT back in 2024, then kept pushing speech tech forward. With the release of GPT-Live, they finally achieved full-duplex capabilities—meaning the AI can listen and speak at the same time. That makes conversations feel a lot more like talking to a real person. In their announcement, OpenAI made their vision clear: in the future, you'll just say what you need, and ChatGPT will quickly advance multiple tasks based on your thoughts.

The standout feature of the new voice mode is that it lets you kick off multiple workflows from a single conversation. While you keep chatting, the AI silently handles tasks in the background. OpenAI gave an example of planning a business trip: you can ask ChatGPT to check your calendar for conflicts, scan your inbox for flight changes, and prepare meeting notes—all while you're making coffee or doing something else. Compared to just using voice as an input box, this "you speak, it handles the task, and both proceed without hindrance" interaction is exactly what GPT-Live aims to differentiate itself with.

Image

Key Points

  • GPT-Live voice mode is now available on the ChatGPT desktop app.
  • Users can start, check, or adjust tasks using voice in Chat, Work, and Codex sections.
  • Full-duplex capability allows simultaneous listening and speaking for more natural conversation.
  • The AI can handle multiple background tasks while users continue talking.