ChatGPT Is Becoming One Workspace, With One Allowance, in December
The Chat and Work tabs are merging, and ChatGPT and Codex will draw from one pool. Here is what to check before your team hits the wall.
By John Martines · November 3, 2026 · 6 min
OpenAI is folding ChatGPT's Chat and Work tabs together and, from December, Plus, Pro and Business seats share one usage allowance with Codex. Who uses what in your office now decides who runs out first.
It's a Tuesday afternoon at a small landscaping company. The office manager has spent the morning in ChatGPT's Work tab, turning a stack of client notes into spring proposals. The owner's nephew, who keeps the website running, has been in Codex fixing the booking form. Both are on the company's Business plan. Around three o'clock, one of them gets a message to wait for a reset.
Today that is unlikely, because each of them is drawing on the kind of usage the other barely touches. From December, it gets more likely: OpenAI says ChatGPT and Codex will share one allowance on Plus, Pro and Business seats. And the screen they work in is changing too.
ChatGPT's desktop app has had three areas: Chat (ask and answer), Work (the agent mode that uses your files and plugins to finish a job, like a deck or a spreadsheet) and Codex (the coding assistant). On November 2, 2026, TechCrunch reported an email OpenAI sent users saying the Chat and Work tabs will merge into one view. Codex keeps its own space.
OpenAI's stated goal is to let one conversation go from exploring an idea to ChatGPT carrying it through: bring in your files and plugins, then ask it to build a website, make a presentation or do things on the web. The merged view goes to a limited group of users for feedback first, before it rolls out widely. No date has been given for everyone.
The second change is about usage. A usage allowance is the amount of AI work your seat includes before you have to wait for it to reset. From December, ChatGPT and Codex will draw on a single, token-based allowance for Plus, Pro and Business users. (A token is the small unit of text the AI reads and writes; heavier models and longer jobs use more of them.) OpenAI says the change won't affect most users, but that, depending on how much they use and which model, some people may need to wait for resets more often.
How it got here:
- September 2026: Anthropic folded its Cowork agent into Claude's main chat, so Claude became one app (we covered it in [Claude Is One App Now](/articles/claude-is-one-app-now)).
- October 29, 2026: after complaints, OpenAI dropped a planned cut to its $200 Pro tier and instead doubled limits on its Pro tiers and Business Premium.
- October 31, 2026: OpenAI's help page on usage limits was updated to describe the shared ChatGPT and Codex allowance coming in December.
- November 2, 2026: OpenAI told users the Chat and Work tabs will merge, starting with a small group.
- December 2026: the shared allowance takes effect for Plus, Pro and Business. OpenAI has not named the day.
Three things change for an office that pays for ChatGPT seats.
The heavy jobs and the everyday jobs now share one tank. Asking a quick question costs little. Asking the agent to build a 20-slide deck from six files, or asking Codex to rework a web page, costs a lot more. Until now, someone who lived in Codex wasn't eating into the same allowance as their ChatGPT work. From December, on these plans, they are.
OpenAI hasn't published the numbers. The help page says use depends on the model and the task, and that you'll see your remaining allowance and reset times under Settings, then Usage. It does not say how big the allowance is per seat. You will learn your team's real limit by watching, not by reading a price sheet.
The screen your staff learned is about to move. If someone wrote down "click Work, then type @ to pick the plugin," that note will go out of date when the merged view reaches your account. Small changes like this are where people quietly stop using a tool.
The upside is real. One place to start a job and finish it is simpler for most people, and OpenAI's record from the last month suggests it listens when paying customers say the limits are too tight.
1. Write down who uses what. Ten minutes. A list of everyone on a ChatGPT seat, with three columns: mostly Chat, uses Work for bigger jobs, uses Codex. In most small offices, one or two names will be in the last two columns. Those are the people to watch in December.
2. Look at Settings, then Usage, in the first week of December. Ask each heavy user to check their remaining allowance and reset time at the end of a busy day. If you are the workspace owner on a Business plan, OpenAI's help pages say you can see workspace credits and download usage reports under Settings, then Billing. One week of real numbers tells you more than any guess.
3. Decide who needs the bigger seat, and only them. If one person keeps hitting resets, upgrade that one seat or buy extra workspace credits where your plan allows it. Don't move the whole office up a tier because of one heavy user. Our guide [Who Needs the Bigger AI Seat](/articles/who-needs-the-bigger-ai-seat) walks through that call.
4. Write a "when you run out" rule. One sentence: "If ChatGPT says to wait for a reset, switch to a lighter model or tell me; never move company work into a personal account." People under deadline will find a workaround. Give them the one you want.
5. Let one person try the merged view first. When it reaches your account, have your AI champion use it for a week, then update your how-to notes and show everyone else in five minutes. Nobody else needs to figure it out alone.
6. Keep coding jobs on one seat. If someone uses Codex for your website or a small internal tool, keep that work on their seat, so a big Codex job doesn't drain the front desk's allowance on a busy day.
- Do you know which seats use Work or Codex for long jobs, by name?
- Has someone on each of those seats found Settings, then Usage, and seen their reset time?
- Does the workspace owner know where the usage reports and extra credits live in Billing?
- Is there a written rule for what staff do when they run out, and does it keep work off personal accounts?
- Are your how-to notes for ChatGPT dated, so you know which ones to redo when the screen changes?
ChatGPT and Claude have now both moved toward one window where you talk, hand over files and let the agent finish the job. The price of that is that AI plans are starting to look like a phone plan: one pool, used at different rates depending on what you ask for. The numbers will keep moving. OpenAI has raised and cut limits within the same month.
You don't need to predict them. You need to know who in your office uses the heavy features, look at the usage screen once a week for the first month, and pay for more only where the numbers show it. That keeps the AI working for the people who need it most, without a surprise reset in the middle of a proposal.
---
Sources (checked November 3, 2026). Chat and Work merge, OpenAI's email, limited rollout first, Codex kept separate, shared allowance for Plus, Pro and Business in December: "OpenAI is finally merging ChatGPT's Chat and Work tabs" (TechCrunch, Nov 2, 2026); also reported by Thurrott and Android Headlines (Nov 2, 2026). Usage depends on model and task, Settings then Usage, help page update: "ChatGPT and Codex Will Share Usage Limits in December" (TechRepublic, Nov 2026) and AlternativeTo (Nov 2026), citing OpenAI's help center page on usage limits (updated Oct 31, 2026). Workspace credits and usage reports for Business owners: OpenAI Help Center, "Using Codex with your ChatGPT plan" (checked Nov 3, 2026). What ChatGPT Work does and which plans have it: OpenAI, "ChatGPT Work for every team" (checked Nov 3, 2026). Pro and Business Premium limits doubled after a planned cut was dropped: "OpenAI doubles ChatGPT Pro usage limits after backlash over planned 50% cut" (BleepingComputer, Oct 2026) and Android Authority (Oct 2026). Claude's Cowork merged into the main chat: Applied AI Insider, "Claude Is One App Now" (2026).