AI Affairs, home

Monday 2 November 2026

Technology

OpenAI plans to merge ChatGPT’s Chat and Work tabs and share Codex usage

A small group will test the combined interface before wider release. Instant use will remain outside the shared allowance, while more demanding tasks will draw on it.

The Pioneer Building in San Francisco, partly screened by trees along the street.
Photo: HaeB, CC BY-SA 4.0, via Wikimedia Commons (cropped)

OpenAI plans to bring ChatGPT’s Chat and Work tabs together and make ChatGPT and Codex draw on a shared usage allowance, according to an email to subscribers reported by TechCrunch on 2 November 2026. The change would let people move from a conversation to a longer task in one place, while making work in the separate Codex coding area count towards the same allowance.

Key points

  • OpenAI plans to test the combined Chat and Work experience with a small group before expanding it.
  • Codex will stay separate in the desktop app but share an allowance with ChatGPT.
  • Unlimited Instant use will not count towards that shared limit.
  • Checkout notices seen by some subscribers point to December for the shared limit.

Chat and Work move into one place

The desktop app currently has ChatGPT and Codex sections, with another choice between Chat and Work inside ChatGPT. OpenAI told subscribers it would begin testing a combined Chat and Work experience with selected users, seek their feedback and then expand it. Codex will remain its own area in the desktop app. The proposed change therefore removes a choice within ChatGPT without folding the coding area into the same screen.

Work is the mode OpenAI introduced in July for longer jobs, including research and coding that takes several steps. Under the proposed interface, a conversation could lead directly into a task: OpenAI says people in the test will be able to supply files and plugins, ask ChatGPT to make a website or presentation, or have it carry out work on the web while they review and guide it. The distinction matters because a longer task involves carrying out work across steps, rather than only returning an answer to a prompt.

Putting together a presentation could start with a question and supplied files, move into a draft in the same place, and leave room to review and guide the work as it proceeds. It would avoid choosing a separate mode before the task begins.

Anthropic brought Claude’s Chat and Cowork functions into one screen in September, TechCrunch reported. OpenAI’s plan follows a similar interface change, though its separate Codex area remains important to anyone who works across conversations and coding tasks. OpenAI has also given Codex reusable cloud environments and security scanning, changes to the coding experience that remain distinct from the Chat and Work interface plan.

ChatGPT and Codex share an allowance

The interface change comes with a change to the subscription counter. OpenAI told subscribers that ChatGPT and Codex would draw from one allowance, according to the email seen by BleepingComputer. OpenAI says unlimited use of Instant is included in the subscription and will sit outside that shared limit. More advanced models and complex tasks will use more of the allowance, depending on the work involved.

The allowance works like a single meter serving several taps: using one of the metered capabilities leaves less available for the others until the allowance resets. That makes the model and the kind of task consequential for someone moving between ChatGPT and Codex, even though the two remain separate places to work. OpenAI says most users should not have to wait for resets more often, but that some may do so depending on their models and usage, TechCrunch reported.

Before this change, Chat, Work and Codex had separate counters, while images, voice and file uploads had additional limits, Trending Topics reported. A shared allowance would change how metered work is divided between ChatGPT and Codex, rather than simply placing their existing counters on one screen. The distinction between Instant and more demanding work also means that a conversation’s effect on the allowance would depend on what it asks the system to do.

OpenAI says subscribers will be able to see how much of the combined allowance remains and when it resets under Settings > Usage. Once it is exhausted, the options described to subscribers are to wait for a reset, move to another subscription or buy credits for eligible features. For developers using Codex alongside ChatGPT, that makes the usage display a way to track capacity across both experiences rather than checking a coding allowance in isolation.

Pro allowances doubled on 29 October

OpenAI doubled the Work and Codex allowances for its Pro 100, Pro 200 and Pro 500 subscriptions, effective 29 October, BleepingComputer reported. That adjustment concerns the existing allowances as OpenAI prepares the shared arrangement. It comes after a difficult introduction for Work: when the mode launched in July, Work and Codex limits were reset twice in one day, and OpenAI added up to 50 percent more usage in August, according to Trending Topics.

OpenAI plans to test the combined interface with a small group in the coming weeks before offering it more widely. OpenAI’s checkout notices seen by some subscribers put the start of the shared ChatGPT and Codex limit in December, Trending Topics reported.

Sources

Topics: Foundation models, Inference