OpenAI Just Gave ChatGPT Desktop A Voice, And 5 Tiers Got It First

	OpenAI Just Gave ChatGPT Desktop A Voice, And 5 Tiers Got It First

OpenAI has added its GPT-Live voice mode to the ChatGPT desktop app, rolling out globally across five subscription plans on macOS and Windows.

Key Points:

  • ChatGPT Voice now runs on macOS and Windows for Plus, Pro, Business, Edu and Enterprise subscribers.
  • The mode drives Chat, Work and Codex threads, and can read the active window on Mac through Appshots.
  • Anthropic shipped its own voice upgrade the same day, adding Opus and Sonnet to Claude.

ChatGPT Voice Reaches macOS And Windows

The company shipped the update Thursday, reaching Plus, Pro, Business, Edu and Enterprise subscribers on both operating systems at once. Release notes clarified that the feature runs inside Chat, Work and Codex threads. Users can open a task by speaking, then ask the assistant to start, check or steer work already running somewhere else.

On macOS, a capability called Appshots lets the app read whichever window a user currently has in focus, alt text included. Voice mode also taps computer use skills, local files and plugins, and it reaches Codex from the iPhone through paired remote access, though the company says Android support has yet to arrive.

Also Read: BitMEX Is Shutting Down After 11 Years With No Buyer In Sight

Guinness Chen Frames Codex Control

Guinness Chen wrote on X that GPT-Live makes conversation feel alive, and that Codex "gives it the power to act." He said he had lived with the feature for several weeks and no longer wants to work any other way. In a demo clip, the company showed a developer opening a thread, filing a pull request and tracing the root cause of a bug through one spoken command.

The same release reworks local projects. A project can now hold several related folders, with one marked as primary for new chats, Git operations and automatic file discovery.

Anthropic Answers With Claude Voice Tools

Anthropic moved on the same day, upgrading Claude's voice mode with its Opus and Sonnet models after months in which the feature ran only on Haiku, its fastest and shallowest tier.

Claude can now reach Gmail, Google Calendar, Slack, Notion and Canva, and it asks permission before touching a tool. Voice conversations count against standard usage limits, and free accounts stay on Haiku with a single connected app.

The two approaches diverge. Anthropic left its voice stack alone, so Claude still listens, pauses, then answers, and the company noted that this release centers on intelligence and tool access rather than the way the words come out.

OpenAI pulled voice mode from its older Mac app in December 2025, and the desktop client went without one for months afterward. GPT-Live landed on phones this month with full-duplex conversation, better interruption handling and CarPlay support, but it was never built to take action on a handset.

Thursday's release is the first time the model has been handed a keyboard, a screen and a file system to work with.

Read Next: HubSpot Fell 12.7% And Analysts Point Straight At OpenAI's New Agent

Disclaimer and Risk Warning: The information provided in this article is for educational and informational purposes only and is based on the author's opinion. It does not constitute financial, investment, legal, or tax advice. Cryptocurrency assets are highly volatile and subject to high risk, including the risk of losing all or a substantial amount of your investment. Trading or holding crypto assets may not be suitable for all investors. The views expressed in this article are solely those of the author(s) and do not represent the official policy or position of Yellow, its founders, or its executives. Always conduct your own thorough research (D.Y.O.R.) and consult a licensed financial professional before making any investment decision.
OpenAI Just Gave ChatGPT Desktop A Voice, And 5 Tiers Got It First | Yellow