ChatGPT and Codex will share one token balance in December
Plus, Pro and Business subscribers lose their separate limits, but OpenAI has not said how many tokens each plan includes.
Subscribers to ChatGPT Plus, Pro or Business who also use Codex, OpenAI's coding assistant, will draw on a single pool of usage starting in December. Whatever one product spends comes out of the other. OpenAI laid out the change in its Help Center, updated Oct. 31.
Usage will be counted in tokens, the chunks of text a model reads and writes. How fast the balance drains depends on the model and the task: the most advanced models and complex jobs cost more. Instant chat does not count against it.
OpenAI ties the change to the arrival of Work in ChatGPT. Those are "agentic" features, meaning the model carries out tasks on its own instead of waiting for each step. The pricing page already says Work shares Codex's prices, credits and limits.
The current setup is different. Plus and standard Business plans get estimates of local messages per five-hour window and can also carry weekly caps, while Pro has no five-hour limit. Under the new scheme, the remaining balance and the reset time will appear together under Settings, in the Usage section. When the balance runs out, users can wait for the reset, change plans or buy credits for the features that accept them.
OpenAI expects most users will not have to "wait" for a reset more often than they do today.
The numbers that would show whether the change is a good deal are missing. The company has not published how many tokens each plan includes, which day in December the switch happens, or what changes for Free, Go, Enterprise and Edu. Without them, the new balance cannot be compared with current limits, and the promise that almost no one will wait longer has no independent test yet.
