Osegi
AI·2 min read

OpenAI to Merge ChatGPT and Codex Usage Limits From December

From December 2026, OpenAI will pool ChatGPT web chat and Codex usage into one shared limit for Plus, Pro and Business subscribers, counted in tokens.

ChatGPT and Codex logos side by side with a single shared usage meter on a dark background.
Image: Хабр

Starting in December 2026, OpenAI will combine the usage limits for ChatGPT web chat and the Codex coding agent into a single shared allowance for Plus, Pro and Business subscribers. Chatting in the browser and working with Codex will draw from the same pool.

Until now, the two services ran on separate limits. A user could hold long conversations in ChatGPT and still have a full Codex quota for programming. Under the new scheme, extended chats can reduce what is left for Codex, and heavy coding sessions can limit how much a user can chat.

Tokens instead of messages

OpenAI will stop counting messages and instead measure the tokens models process and generate. Consumption depends on the chosen model and the complexity of the task, so one long conversation with a resource-heavy model could cost far more than dozens of simple prompts.

Instant mode is exempt: ordinary quick replies will not count against the shared limit. The company has not yet said how its reasoning modes, including Medium, High and Extra High, will be counted.

Users will be able to track their balance under Settings → Usage, which will show the remaining allowance and when it resets. Once the limit is exhausted, OpenAI will offer to wait for the reset, move to a more expensive plan or buy additional credits.

Why OpenAI is doing it

The company says the change reflects the gradual move of agent capabilities from ChatGPT Work into the main ChatGPT interface. It promises that most users will not have to wait for limits to recover more often than they do now.

OpenAI has not disclosed the size of the new quotas, the exact December date or the detailed consumption rules, so the impact on a typical $20-a-month Plus subscriber cannot yet be judged.

The announcement follows a move on October 29, when OpenAI raised Work and Codex limits for some Pro and Business Premium plans, including a doubling for Pro 100 and Pro 500. Plus received no increase. Plus users who rely on both ChatGPT for long discussions and Codex for development may lose the benefit of two independent limits. Whether the shared pool is large enough will be clear only once the new quotas are published.

Tags:openaichatgptcodexusage limitssubscriptions

More from this day

See all →

More in: AI

See all →