ChatGPT and Codex Share One Usage Allowance From December
OpenAI says that from December 2026, Plus, Pro and Business plans will use one token-based usage allowance across ChatGPT and Codex. Heavy coding sessions will draw on the same limit as ordinary ChatGPT work, and Instant chat is excluded from the shared pool. OpenAI has not published an exact December start date or the size of each plan's allowance. For a business that runs automated work through Codex or ChatGPT Work, one budget will now cover both kinds of task. Before the change takes effect, teams should test their typical workloads and watch the combined balance in Settings under Usage.
In short
- From December 2026, Plus, Pro and Business plans share one token allowance across ChatGPT and Codex, according to OpenAI's help article.
- Instant chat is excluded from the shared pool, but agentic work such as Codex and ChatGPT Work is included.
- OpenAI has not published a December start day, per-plan token amounts or any change to the five-hour and weekly windows.
- Background automations and staff work will compete for the same balance, so track usage in Settings under Usage before the change.
What happened
According to OpenAI's Help Center article on ChatGPT and Codex usage limits, the usage allowance for Plus, Pro and Business plans will be shared across ChatGPT and Codex starting in December 2026. The article was last updated on October 31, 2026. OpenAI ties the change to bringing the agentic features of ChatGPT Work into the main ChatGPT experience. The change applies to OpenAI's paid plans worldwide, and the help page does not limit it to a single market.
Windows Report, a technology news site, reported a checkout notice that some subscribers had seen. The notice used similar wording about a shared limit across ChatGPT and Codex starting in December. Windows Report also noted that OpenAI has not publicly confirmed whether every subscription tier will receive the shared limit. The help article is the primary source, and the notice is a secondary report.
What the shared pool covers
Usage on both products will be counted in tokens. A heavier model or a harder task uses more of the allowance, so agent-heavy Codex work will draw on the same budget as regular chat. Instant chat does not count against the pool. Remaining usage and reset times will appear in Settings under Usage.
OpenAI already groups several work features under one balance. Its Codex pricing page states that ChatGPT Work usage inside ChatGPT follows the same prices, credits and limits as Codex. Its ChatGPT Business release notes say that Business plans include usage for Workspace Agents, ChatGPT for Excel and ChatGPT for PowerPoint through the general Codex agentic usage pool. The December change widens that existing pool to cover ChatGPT more broadly.
What is not yet known
The help article gives no calendar day in December. It also publishes no token amount for any plan. It does not say whether the current five-hour and weekly usage windows will change. These are the three figures a budget owner needs most, and none is available yet.
Why it matters for businesses that automate work
Many teams that use AI for routine work now run two kinds of usage at once. One is conversational work, such as drafting a reply or summarising a file. The other is agentic work, such as a Codex task that checks records or a ChatGPT Work task that builds a report. Under the new arrangement, both kinds of work will draw from one balance. A workflow that runs in the background can therefore use up the allowance that staff rely on during the working day.
What changes in practice
- One balance replaces separate views. Owners and operations leads should check Settings under Usage once the change is live, and record the reset times.
- Background automations and staff use compete for the same tokens. A scheduled report run can reduce the room left for a person who needs Codex that afternoon.
- Instant chat stays outside the pool. Quick factual questions in chat do not draw from the shared allowance, according to the help article.
- Limits have a defined fallback. At the limit, users can wait for a reset, upgrade the plan, or buy credits for eligible features, as OpenAI lists in the same article.
A sole owner using one Plus seat will see the change as one meter instead of two. A larger team on Business should assign automations to named accounts and track which workflows use the most tokens. In both cases, the first step is the same: run the tasks you rely on now and note how much of the allowance each one uses. OpenAI has not published the token sizes, so that measurement is the only reliable way to plan.
Limits of this change
The pool is a usage budget. It does not guarantee capacity for any particular task, and a workflow that reaches the limit stops until the reset, an upgrade or a credit purchase. Automation also does not remove the need for a person to check outputs before they reach a customer, a ledger or a tax filing.
What to watch next
Three things will settle the practical impact. The first is the exact start day in December. The second is the token allowance for each of the Plus, Pro and Business plans. The third is whether the five-hour and weekly windows stay in place. Business owners should also watch whether OpenAI confirms the change for every subscription tier, since Windows Report notes that this has not yet been stated publicly.
Sources
How AiStaffo would automate this
AiStaffo would start by listing every automated task that touches ChatGPT or Codex, such as a reconciliation run, a follow-up draft or a monthly report build, and recording which account runs each one. The system would log the token use of each task against the Usage view, so the owner sees which workflows take the largest share of the pool. Routine runs would be scheduled for off-peak windows where the business allows it, and a person would still review anything that goes to a customer, a supplier or a ledger. The owner decides which tasks keep priority when the allowance runs low. To see how this would work with your own workload, book a free automation audit.
Questions people ask
When does the shared ChatGPT and Codex limit start?
Does normal ChatGPT chat use up the Codex allowance?
How many tokens does each plan get?
Does the change apply to every ChatGPT plan?
What happens when a business reaches the limit?
Book a free automation audit
Thirty minutes. We look at one process you run every week and tell you exactly what an AI worker would take off your desk, and what it would not.
