OpenAI’s new voice mode makes it to the ChatGPT desktop app

OpenAI updated its ChatGPT desktop app on Thursday to support ChatGPT Voice, letting users talk to the app to control AI agents and perform multi-step tasks on their computer rather than typing commands. The feature runs on GPT-Live, the voice model family OpenAI launched earlier this month, and works with both ChatGPT Work and Codex, tapping computer-use skills to look up websites and apps, with macOS users able to grant screen access through Appshots so the assistant can see on-screen context including alt text. In a demo, OpenAI showed a developer asking ChatGPT to create a new thread, open a pull request, and find the root cause of a bug using a single spoken command, illustrating the jump in capability from the original smartphone version of ChatGPT Voice, which offered smoother conversation and better interruption handling but could not take action on a device at all. The rollout lands the same week Anthropic separately updated Claude’s voice mode with support for its Opus, Sonnet, and Haiku models and integrations with apps like Gmail, Calendar, and Notion, intensifying a race between the two companies to make voice the primary way people direct AI agents rather than just chat with them. Where OpenAI’s update emphasizes acting directly on a user’s computer, Anthropic’s leans on connecting to a wider set of everyday productivity apps, a split that suggests the two companies see somewhat different futures for how voice-driven AI assistants should actually get work done.