OpenAI will retire its legacy ChatGPT Classic Mac app on 30 January 2027, the company told users on Monday, ending support for its original native macOS software as it finishes moving desktop users to a unified client that combines conversational chat, coding and workplace tools.
Key points
- OpenAI set a formal shutdown date of 30 January 2027 for the legacy Mac application.
- The company is directing users to download its unified Electron-based desktop application.
- A coming update will merge the separate Chat and Work tabs into a single conversational view.
- Subscribers will move to a pooled usage allowance shared across general chat and Codex programming tasks.
Retirement date set for legacy client
Users opening the older Mac software on Monday were met with an in-app notice setting the end date. The banner states that the client will stop working on 30 January 2027, advising users that their full conversation history will transfer automatically when they switch to the modern desktop build available on the official ChatGPT downloads page.
The move draws a line under a transition that started several months ago. When OpenAI launched its multi-purpose desktop application, it rebranded the original client as ChatGPT Classic. That original software had been built specifically for macOS, but OpenAI left it largely untouched after releasing its successor.
Over recent months, the older build fell behind. It lacked controls for newer models, including the option to adjust reasoning effort levels, and suffered from growing reliability problems as backend systems shifted towards features designed for the newer software.
The shift away from native macOS software
The retirement confirms OpenAI will not maintain a separate, native Mac build. The company faced criticism from Mac users earlier this year when it replaced the swift, lightweight Swift-based client with a cross-platform version built on the Electron framework.
Desktop users voiced frustration with both performance and interface design. The unified software initially placed standard chat conversations alongside Work and Codex workspaces in a layout that made simple chats difficult to find. OpenAI reacted by adding toggles across the top bar and sidebar shortcuts, but the older client remained popular among users who wanted a plain chat window without developer tooling.
The company also adjusted commercial terms alongside these software changes. As previously reported, OpenAI sets ChatGPT subscription limits on payment cards and devices to control how accounts run across multiple desktop installations and shared hardware.
Merging Chat and Work into one interface
Alongside the retirement notice, OpenAI told subscribers by email that it will begin testing a combined Chat and Work interface in the coming weeks. The test will start with a limited group of accounts before rolling out to all users.
The change aims to fix the awkward separation inside the current desktop build. Greg Brockman, the president of OpenAI, previously acknowledged to users that the multi-tab layout was “kind of a mess” and pledged to drop the dedicated Work tab before the end of the year.
Under the new layout, users will handle standard questions and multi-step project tasks in the same prompt box. Users can bring local files and external plugins into regular chats, allowing the model to draft presentations, write code or browse web resources without forcing the user to switch to a different workspace. Codex will remain in a dedicated panel designed for programmers.
Deep software integration has become a major focus for the desktop team. Earlier this year, OpenAI connects ChatGPT to Apple iCloud Mail and Calendar to let desktop models coordinate directly with local scheduling tools, reducing the need for standalone enterprise tabs.
New shared allowances for desktop tasks
Bringing those features together also alters how OpenAI calculates usage. Instead of offering separate rate limits for conversational models and developer workspaces, the company plans to introduce a shared allowance across ChatGPT and Codex.
OpenAI told customers that standard interactions through the Unlimited Instant tier will not consume allowance units. Instead, heavier queries using deep reasoning or autonomous web actions will draw from the central quota based on model size and processing time. The company stated that most users should not encounter more frequent rate resets under the combined pool.
The policy follows broader adjustments to paid tiers. Last month, ChatGPT Pro limits doubled as OpenAI combines Chat and Work, giving top-tier accounts higher query caps to prepare for automated multi-step desktop tasks.
Mac users have roughly three months to migrate their workflows to the new software before the legacy build stops connecting to OpenAI servers on 30 January 2027.
