From December, OpenAI will combine ChatGPT’s Chat and Work modes and give Plus, Pro and Business users a shared, token-based quota for ChatGPT and Codex. More complex models will use more of that allowance, while simple instant chats will not count, according to The Decoder.
OpenAI Watch analysis
What happened
The change replaces the current all-you-can-prompt arrangement for those plans with a common quota across the two products. The Decoder says the new system is due to start in December. Its report does not provide the size of the quotas or further detail on how usage will be measured.
That means the practical cost of a session will depend partly on which model a user chooses: complex models draw more from the shared allowance, while simple instant chats are excluded. The report’s suggestion that heavy users on cheaper plans could see a drop in quality is an inference, not a confirmed OpenAI policy.
Why it matters
A shared pool makes ChatGPT and Codex less like separate destinations and more like two ways to spend from the same monthly budget. For people who use both, a long coding session could affect how much room remains for other ChatGPT work. The missing quota sizes make it hard to judge the impact yet.
Model choice may also become a budget decision. Users could find themselves weighing a more capable model against how much of their allowance it consumes, rather than simply choosing the best tool for the task.
Our read
This is a meaningful shift in how paid access is rationed, not proof that every user will get less. The details that matter most are still to come: quota sizes, reset periods and what happens when users reach their limit. Until those are clear, the headline change is the shared meter, not a guaranteed downgrade.
What to watch
- The quota sizes and reset periods for Plus, Pro and Business.
- How OpenAI defines the usage cost of different models and tasks.
- What users can do when they exhaust the shared allowance.
Discussion spark: Should ChatGPT and Codex draw from one shared allowance, or should paid users get separate budgets for each product?
Sources and evidence
- OpenAI's upcoming chat-and-work merger in ChatGPT signals the end of all-you-can-prompt pricing (1 November 2026, 09:51 UTC)
OpenAI Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by OpenAI.