OpenAI to merge ChatGPT and Codex usage into one token allowance from December

What OpenAI changed

OpenAI updated its help center on 31 October 2026 with a change to how paid plans are metered. From December 2026, the usage allowance on ChatGPT Plus, Pro and Business will be shared across ChatGPT and Codex, according to the help center wording quoted by several outlets. Usage in both products will be counted in tokens, so a heavier model or a harder task draws down more of the allowance. Chats in Instant mode do not count against it.

Today, OpenAI's help article on Work and Codex usage already describes one allowance shared by ChatGPT Work and Codex, with five-hour and weekly windows, and separate message limits for GPT-6 Pro in Chat. The December change folds regular ChatGPT use into that same pool. Remaining usage and reset times will appear in one place, under Settings, Usage. Under the current rules a new window starts with the first message in Work or Codex after the previous window ends, banked full resets refresh both the five-hour and the weekly limit, and purchased resets apply as soon as checkout succeeds. None of that is described for the combined pool yet.

Much is missing. The help center gives no calendar day in December and publishes no token amount for any plan. It also does not say whether the five-hour and weekly windows stay, change or disappear once chat and Codex share a budget. OpenAI pages are not yet consistent on this: the current Work and Codex article says Pro 100, Pro 200 and Pro 500 have no five-hour limit but still have a weekly allowance, a detail the December note does not address.

A week of changes to limits

The update follows a turbulent few days for OpenAI's top plans. Pro subscribers had been told that the Work and Codex allowance on Pro 200 would fall from 20 times the Plus level to 10 times from 30 October. On 29 October, Tibo, who leads Codex and ChatGPT Work at OpenAI, said on X that the cut was cancelled and that usage limits on the $100, $200 and $500 Pro tiers and on Business Premium were doubled with immediate effect, as reported by gHacks and Android Authority. Outlets differ on whether the doubling covers every limit or only the Work and Codex allowance, and OpenAI does not publish numeric limits that would settle it.

Enterprise and Edu are not named in the December change. OpenAI's rate card says Enterprise and Edu workspaces on flexible, credit-based pricing have no fixed rate limits and scale usage with credits, and that Codex, ChatGPT Work and workspace agents share agentic usage and credits where a plan includes them. Free and Go plans are not mentioned either.

Competitors are moving in a similar direction. Anthropic renamed its Claude Max plans on 30 October: the $100 tier is now Max and the $200 tier is now Ultra, and verified Ultra accounts no longer have a five-hour limit, leaving the weekly cap as the only one, according to The Next Web and Windows Report.

What it means for subscribers

For people who mostly chat, the change may be invisible, because Instant chat is excluded. For people who run long Codex sessions or agent tasks in ChatGPT Work and also use reasoning models in chat, the two will compete for one budget from December. A heavy coding day could leave less room for research in ChatGPT, and the reverse. Pro, which OpenAI sells at $100, $200 or $500 a month, has the largest allowance, so the practical effect will depend on whether OpenAI sizes the shared pool to cover what each tier used to get from chat and Codex combined, or something smaller.

A shared token budget is simpler to explain than separate counters per product and model, and it lets OpenAI price expensive models such as GPT-6 Astra by what they consume rather than by message count. The cost of that simplicity is predictability: without published token amounts, subscribers cannot work out in advance how many tasks a month their plan covers.

Things to watch before December: the exact start date, whether OpenAI publishes allowance sizes for Plus, Pro and Business, how the five-hour and weekly windows are applied to the combined pool, and whether Enterprise, Edu, Free and Go move to the same model. Teams that budget Codex work around current limits should check Settings, Usage once the change goes live.

Questions people ask

When do ChatGPT and Codex start sharing one usage limit?

OpenAI says December 2026 but has not given an exact day. The change covers Plus, Pro and Business plans.

Does Instant chat use the shared allowance?

No. OpenAI's help center says Instant chat does not count against the shared ChatGPT and Codex allowance.

How big is the new allowance?

OpenAI has not published token amounts for any plan. Remaining usage and reset times will be shown under Settings, Usage.

Sources

1. Managing usage with GPT-6 Astra in Work and Codex (OpenAI, 31 Oct 2026)
2. ChatGPT Rate Card (Business, Enterprise/Edu credit-based pricing) (OpenAI, 31 Oct 2026)
3. ChatGPT and Codex will share one usage allowance from December (DEV Community, 31 Oct 2026)
4. OpenAI Confirms One Shared Usage Limit for ChatGPT and Codex From December (Pasquale Pillitteri, 1 Nov 2026)
5. OpenAI Doubles ChatGPT Pro and Business Premium Usage Limits After Dropping a Planned Pro 200 Cut (gHacks, 30 Oct 2026)
6. OpenAI caves to backlash and reverses course on reducing ChatGPT usage limits (Android Authority, 30 Oct 2026)
7. Anthropic renames Claude Max plans and drops Ultra's five-hour limit (The Next Web, 30 Oct 2026)