Guide · checked 2 November 2026

ChatGPT and Codex are moving to one shared usage limit in December: what OpenAI says, and what it doesn't

Independent project. Not affiliated with Anthropic or OpenAI.

If you use Codex for coding and ChatGPT for everything else, your usage is about to land in one bucket. This page sticks to what OpenAI's Help Center actually says, marks what is not confirmed, and suggests a few habits that make a shared limit less of a surprise.

The short answer

OpenAI's Help Center says that starting in December, usage allowance on Plus, Pro, and Business plans will be shared across ChatGPT and Codex.
It does not give an exact date, the size of the shared allowance, or how it compares with today's.
If you hit a limit mid-task, you will want a way to hand work over. Limit Rescue is a small local command line tool that warns you at 80% and 95% of your usage limit, writes a "Where I left off" note, and lets you carry on in the other agent with limitrescue resume codex or limitrescue resume claude. It works with Claude Code and Codex. It is in early access, not public yet.

Join the list

Limit Rescue v0.1.0 is not on npm yet, is not public yet, and has only been tested on macOS so far. We have not tested it against OpenAI's December change, because that change has not happened yet. We will update this page when we know more.

What OpenAI's Help Center says

The primary source is OpenAI's article Understanding ChatGPT and Codex usage limits. It was marked "Updated: 20 hours ago" when we read it on 2 November 2026. Everything in this table comes from that page.

Claim	What the page says
Timing	"Starting in December"
Plans named	Plus, Pro, and Business
What is shared	The plan's usage allowance, across ChatGPT and Codex
Why	The page says it is "as we bring the agentic capabilities of Work into the main ChatGPT experience"
How usage is measured	In tokens, "the units of information a model processes and generates"
What affects how much you use	The model you choose and the task you ask it to do. More advanced models and complex tasks generally use more of your allowance
Instant chat	"Instant chat will not draw from your usage"
Where to look	You will be able to see your combined remaining usage and reset times in Settings, then Usage
At the limit	You can wait for it to reset, upgrade your plan, or buy credits to continue using eligible features
OpenAI's expectation	"We expect most users won't need to wait for their usage to reset more often than they do now"

The exact opening sentence is: "Starting in December, Plus, Pro, and Business plans’ usage allowance will be shared across ChatGPT and Codex as we bring the agentic capabilities of Work into the main ChatGPT experience."

The page also says OpenAI is "preparing carefully for this transition to help prevent disruption to your workflows".

What is already true today

Two other OpenAI Help Center pages describe the current setup. They are separate from the December change.

Using Codex with your ChatGPT plan says that for plans that use shared allowances or credits, Codex, ChatGPT Work, ChatGPT for Excel, and Workspace Agents use a shared allowance and credit pool when those features are available on your plan. So the idea of Codex sharing a pool with other agentic features is not new. What changes in December is that the page says the allowance will be shared across ChatGPT and Codex.
The same page says that if you are close to a Codex limit you can open Settings or your usage dashboard, and that in an active Codex CLI session you can enter /status.
It also says that starting a new Codex chat does not reset your usage limits, and that OpenAI Support does not reset ChatGPT or Codex usage limits.
If you reach a usage limit during an active turn on a plan with included allowances or credit-based billing, the page says Codex can continue working on that turn, subject to fair-use limits. After that turn, the limit banner shows your options, such as adding credits, applying an available reset, upgrading, or waiting.
OpenAI's article on using credits for flexible usage in ChatGPT personal plans covers buying credits.

What is not confirmed

None of these are answered on the pages we read. Please treat them as open questions, not as facts.

The exact date in December. The page says "Starting in December" and nothing more specific.
How big the shared allowance is. The page does not give token amounts for any plan.
Whether the shared allowance is bigger, smaller, or the same as the current allowances combined. The page says OpenAI expects most people will not wait more often than now, which is an expectation, not a number.
Whether the five-hour and weekly windows change. The December article does not mention windows. The other Codex pages we read describe the windows as they work today.
Whether other plans are affected. The page names Plus, Pro, and Business. It does not say what happens on Free, Go, Enterprise, or Education plans, and the Codex page we read says Codex is included across plans, including Free and Go.
How Codex in the CLI will show the combined number. The page says Settings, then Usage will show it. It does not say what /status will show after the change.
How the allowance is split between chat and agentic work beyond "Instant chat will not draw from your usage".

What the press says

The New Stack published an article on 2 November 2026 about the change. It is a news report, not an OpenAI source, so we use it only for context: it reports that OpenAI declined to give the exact December date or say whether the combined allowance equals the existing separate allowances. That matches what we see missing from the Help Center page. We have not used any other claim from the press.

What to do now

You do not need to change anything yet. A few habits will help whichever way the numbers land.

Know where to look. Today, check Settings or your usage dashboard for Codex, and /status inside the Codex CLI. After the change, OpenAI says Settings, then Usage will show combined remaining usage and reset times.
Expect heavy chat and heavy coding to draw from the same place. The page says the allowance depends on the model and the task, and that more advanced models and complex tasks use more.
Keep your work easy to hand over. Commit often. Keep a short note of what is done and what is next. Then a limit, on any tool, is a pause and not a loss.
Do not rely on starting a fresh chat. The Codex page says it does not reset your limits.

Quick answers

OpenAI's Help Center says "Starting in December" for Plus, Pro, and Business plans. It does not give an exact date.

How big will the shared allowance be?

The Help Center page does not say. It gives no token amounts and does not compare the shared allowance with today's separate ones. OpenAI says it expects most users will not need to wait for a reset more often than they do now.

Where will I see my combined usage?

OpenAI says you will be able to see combined remaining usage and reset times in Settings, then Usage. Today, check Settings or your usage dashboard for Codex, and enter /status in an active Codex CLI session.

Will Limit Rescue warnings match the new numbers?

We do not know yet. The December change has not happened, so Limit Rescue has not been tested against it. Limit Rescue v0.1.0 warns at 80% and 95% of your usage limit for Claude Code and Codex, and it has only been tested on macOS so far.

Where Limit Rescue helps

Limit Rescue sits next to your coding agents and does three small things:

It warns you at 80% and 95% of your usage limit.
It writes a "Where I left off" note so another agent can pick up the work.
limitrescue resume codex or limitrescue resume claude continues the work in the other agent.

It runs on your machine and makes no network calls. It supports Claude Code and Codex only, and v0.1.0 has been tested on macOS only. It is not on npm yet and is not public yet. We do not know how Codex will report usage after the December change, so we will not promise that Limit Rescue's warnings match the new Settings, then Usage numbers until we have tested it.

If you also use Claude Code, see what to do when Claude Code hits its limit mid-task and how to switch from Claude Code to Codex without losing your place.