One pool for ChatGPT and Codex: OpenAI merges your allowance in December — and hasn't told you how big it is

Starting in December, your ChatGPT chats and your Codex coding sessions come out of the same bucket. OpenAI's Help Center article "Understanding ChatGPT and Codex usage limits," updated October 31, 2026, says it plainly: "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." There is no blog post and no X announcement — the change arrived as a help-page edit, spotted October 30 and reported November 1–2. [OpenAI Help Center, updated Oct 31, 2026, as reported by Unite.AI, CellCog, and Derrington Collaborative AI, Nov 1–2, 2026]

The headline numbers

That last row is the story. OpenAI has moved the meter to tokens without publishing the size of the meter. Until the December rollout shows your combined balance in Settings → Usage, no subscriber can calculate whether the shared pool is bigger or smaller than the two separate limits it replaces.

What actually changes

This is a consolidation, not a first step. On Plus and Pro, Codex, ChatGPT Work, Excel, and PowerPoint already share one agentic allowance; Business seats on credit-based pricing already pay token rates for Work and Codex under OpenAI's rate card. What December adds is advanced-model chat: the conversations you have in ChatGPT will draw from the same pool as the coding work Codex does for you. Only the fast default — Instant chat — sits outside it.

flowchart TD
    A["Today<br/>ChatGPT advanced-model chat: its own limits<br/>Codex / Work / Excel / PowerPoint: shared agentic allowance"]
    A --> B["December<br/>One token-based allowance across ChatGPT and Codex"]
    B --> C["Instant chat: exempt<br/>Draws nothing from the pool"]
    B --> D["At the limit: wait for reset,<br/>upgrade your plan, or buy credits"]
    B --> E["Settings → Usage shows<br/>combined balance and reset times"]

The "buy credits" option is the hinge. OpenAI offers purchased usage credits for eligible features after the included allowance is used. Separately, on October 28 it announced monthly API credits equal to each eligible Pro tier's price ($100, $200, $500), and on October 29 it doubled the Work and Codex allowances on Pro 100 and Pro 500 while cancelling the planned Pro 200 cut. The subscription API credits are not usage top-ups: they apply to eligible API usage after activation and do not increase ChatGPT or signed-in Codex limits. December's shared pool brings chat and Codex under one allowance, with paid usage credits available for eligible features at the edge.

The per-unit math nobody can publish yet

Here is the price ladder the pool sits on, with the figures OpenAI has actually published:

What cannot be computed — and what every other teardown will pretend to compute — is dollars per token. Without a published token allowance per plan, there is no $/token figure, no "this is a cut" and no "this is a raise." Anyone telling you the shared pool is bigger or smaller than today's separate limits is guessing. OpenAI says it expects most users "won't need to wait for their usage to reset more often than they do now." That is a prediction, not a number.

flowchart LR
    L["Light user<br/>Mostly Instant chat + occasional Pro chat<br/>→ Likely unaffected"]
    H["Heavy dual user<br/>Long ChatGPT sessions AND Codex runs<br/>→ The pool is now one budget, not two"]
    T["Team on Business seats<br/>Chat + Codex across seats<br/>→ Audit per-seat usage before December"]
    L --> H --> T

What it means for you

If you're a light user — mostly Instant chat with occasional advanced-model questions — this likely changes nothing. Instant chat is exempt, and OpenAI's own expectation is that most users won't hit resets more often. Don't upgrade on anxiety.

If you're a heavy dual user — long ChatGPT sessions and Codex runs — you just lost a wall between two budgets. A week that previously ran chat to its limit while Codex still had room (or vice versa) now drains one pool from both ends. Start checking Settings → Usage weekly now, so that when December lands you can see whether your combined burn rate fits the pool — and which side is eating it.

If you run a team on Business seats: audit per-seat usage before December. The at-limit menu — wait, upgrade, or buy credits — means overage now has a price even if OpenAI hasn't published the per-plan token figures yet. Decide in advance which seats get credits and which get capped by waiting; don't let the first shared-pool exhaustion make that decision for you in the middle of a sprint.

The pattern to watch

Three OpenAI moves in five days tell one story: October 28, API credits equal to each Pro tier's price; October 29, doubled Work and Codex allowances and a cancelled cut; October 31, one shared token pool from December. ChatGPT and Codex are converging on a shared usage allowance, with purchased credits for eligible features. The separately claimed subscription API credit benefit remains a different balance. The flat-rate era isn't ending with a price rise. It's ending with a meter.

Watch the December rollout for the one number that matters: the token allowance per plan, as shown in Settings → Usage. Until that appears, treat every claim about the pool's size — including OpenAI's reassurance — as unverified.

Every figure dated, every claim sourced. The AI Bill is an independent practice — no vendor sponsorship, no affiliate links, no paid placement.

Sources: OpenAI Help Center, Understanding ChatGPT and Codex usage limits; OpenAI, API credits for ChatGPT subscribers; OpenAI Help Center, About ChatGPT Pro tiers. Related reporting: Unite.AI, CellCog, Derrington Collaborative AI, and 7minAI. Related The AI Bill essays: API credits and Pro allowance changes.

Next: Book a free 45-minute teardown — screen-share of your last 3 months of AI spend, fixed-fee quote before anything starts.