OpenAI has told subscribers that starting in December, the usage allowance on Plus, Pro, and Business plans will be shared across ChatGPT and Codex. The same update leaves out the one number anyone would need to budget for it: how big that allowance is.
If your team already leans on AI subscriptions for daily work, that gap matters more than the headline. Our view is simple. Don't build next quarter's AI budget on this change yet. Measure what your team actually uses this month, because that is the figure you will hold up against the new allowance once OpenAI publishes it.
What OpenAI has actually said
The wording below is OpenAI's help center text as quoted by Beri.net, OrcaRouter, and DataNorth in the first days of November 2026.
- Who and when. "Starting in December, Plus, Pro, and Business plans' usage allowance will be shared across ChatGPT and Codex."
- The unit. Usage is measured in tokens, not in messages.
- What is exempt. "Instant chat will not draw from your usage." That is not the same as unlimited, and OpenAI's text does not claim it is.
- What costs more. "More advanced models and complex tasks generally use more of your allowance."
- Where you see it. Settings, then Usage, will show your combined remaining usage and when it resets.
- At the limit. You can wait for the reset, upgrade, or buy credits for eligible features.
- The reassurance. OpenAI says it expects most users won't need to wait for a reset more often than they do now. It publishes no figures to back that up.

What is still missing
Four pieces are not public, according to the same coverage. The allowance size for each plan. The exact December start date. How chat tokens convert into the shared allowance. And whether today's five-hour window and weekly limits carry over: Plus and Business Standard currently have a five-hour window with possible weekly limits on top, while Pro has no five-hour limit.
The help article also names only Plus, Pro, and Business. It does not say anything about Free, Go, Enterprise, or Edu.
So the honest summary is this: we know the shape of the change and none of its size.
Why tokens change the arithmetic
A message count is easy to forecast. Ten people, a few dozen messages a day each, done. Tokens are not like that. A quick question in chat and a long coding task that works through many files are both one request, yet they are very different in size. OpenAI says as much: advanced models and complex tasks draw more.
That means two people on the same plan can run out on different days. Per-seat budgeting assumes every seat behaves alike. A shared token allowance punishes that assumption, because one person's chat and coding now draw from the same meter. Someone who uses Codex for long tasks and ChatGPT for research can burn through the allowance for both with a single habit.

The matrix is our reading of OpenAI's statement, not a published rate. Treat it as a direction, not a calculation.
What to do before the numbers arrive
You do not need a tool for this. A shared spreadsheet is enough, as long as you start now and keep it honest.

Flag the cutoffs especially. A team that hits the limit twice a week is in a different position from one that never has, and no pricing page will tell you which one you are.
When the limit hits anyway
OpenAI lists three responses. None is wrong, but with no allowance sizes published, two of them cannot be judged yet.

Upgrading or buying credits today means paying before you know what you are buying. If you are blocked right now, that may still be the right call. If you are not, waiting a few weeks costs you very little.
Our take
Nothing here says the change will be worse for you. OpenAI says most users should not wait more often than they do now, and that may well hold. The point is that nobody outside OpenAI can check it yet.
When the sizes are published, a team with a month of usage notes can answer in an afternoon whether its current plan still fits. A team without notes will find out the hard way, in the middle of a sprint, when the allowance runs dry. That is the difference a spreadsheet makes.
If you want a second pair of eyes on how your team uses AI tools, reach us at bisnis@xetup.id or WhatsApp +62 822 6515 7644.
Sources: Beri.net, 1 November 2026; OrcaRouter, 2 November 2026; DataNorth, 2 November 2026; FourWeekMBA, 31 October 2026 (updated 5 November).
