One pool, two products
OpenAI has told paying subscribers that from December their usage allowance will be shared between ChatGPT and Codex, its coding agent. The change is set out in a help centre article updated on 1 November, which says the shared allowance applies to the Plus, Pro and Business plans and arrives “as we bring the agentic capabilities of Work into the main ChatGPT experience”.
Until now the two have been metered separately. After the change, a long Codex run and an afternoon of ChatGPT prompting draw down the same balance.
The unit of account becomes the token
The second change in the same paragraph matters as much as the first. “Your usage for ChatGPT and Codex will be measured in tokens,” OpenAI writes, “the units of information a model processes and generates.”
That replaces a limit a subscriber could reason about — a number of messages, a number of runs — with one that varies by the thing being asked. OpenAI says how much of the allowance a request consumes “will depend on the model you choose and the task you ask it to do”, and that “more advanced models and complex tasks generally use more”.
A coding agent that reads a repository, runs tests and iterates is, in token terms, nothing like a question typed into a chat box. Under a shared token pool, the expensive half of the pair is the one that sets how much of the month is left.
What stays free of the meter
One carve-out is named. “Instant chat will not draw from your usage,” the article says — ordinary quick conversation is outside the allowance. Everything else, including model choice and agentic work, is inside it.
OpenAI says subscribers will be able to see a combined remaining balance and reset times under Settings → Usage, and that anyone who hits the limit can wait for the reset, move up a plan, or buy credits to keep using eligible features.
The reassurance, and what it leaves open
The company says it is “preparing carefully for this transition to help prevent disruption to your workflows” and expects that “most users won’t need to wait for their usage to reset more often than they do now”. That is a statement of expectation rather than a commitment, and it is the only guidance the page offers about how large the pooled allowance will be.
No token figures are published for any plan. Nothing on the page says how many tokens a Plus, Pro or Business subscription will include, how the pool resets, or what the credits cost. A subscriber cannot, from this document, work out whether their current pattern of use fits inside the new limit.
What to watch
The numbers are the story, and they are not out yet. Watch for the per-plan token allowances, the price of overage credits, and whether the Settings → Usage view ships before December rather than alongside the change. Until those land, the only certain thing is that Codex work and ChatGPT work will compete for the same budget.
