ChatGPT and Codex Are Set to Share Usage Limits Starting in December

XMLans Posted on 1 days ago 14 Views


OpenAI says that, starting in December, usage allowances on ChatGPT Plus, Pro and Business will be shared across ChatGPT and Codex as Work's agent capabilities come into the main ChatGPT experience. That means some ChatGPT activity and Codex will draw from the same allowance. There is an important exception: Instant chat will not draw from that usage, according to the OpenAI Help Center.

Screenshot accompanying the ChatGPT and Codex usage update

What changes?

Currently, Work and Codex already share an allowance, while some Chat limits are separate. OpenAI says the combined usage will be measured in tokens, and the amount consumed will depend on the model and task. More capable models and complex tasks generally use more of the allowance. The announcement does not give a specific December start date, exact token rates, or a complete list of which future chat modes draw from the pool; it explicitly excludes Instant chat.

The company says users will be able to see their combined remaining usage and reset times in Settings → Usage. At the limit, users can wait for a reset or, where eligible, buy credits or upgrade. Those options and their costs depend on the plan and feature.

My concern as a Plus user

Before this change, although Plus users could not use the Pro model in Chat, I personally found ordinary GPT-6 High conversations very generous: on one day I sent well over 100 messages without hitting a limit. That is my experience, not a promise that such usage is unlimited for everyone.

If more of my conversations share the same pool as Codex, I may need to budget that allowance more carefully. I worry that this could effectively reduce the amount of work I can do, although the real impact will depend on the final allowance and usage rates. OpenAI says it expects most users will not need to wait for resets more often than they do now; that remains a company expectation rather than a guarantee.

Different reasoning settings may consume different amounts, but the December announcement does not publish specific multipliers. It also does not say exactly which models and reasoning controls will appear in the future Chat interface. We'll have to wait for those details before comparing the change directly with other services.

Translated from the original Chinese article, published October 31, 2026. The author's experience and concerns are retained; the Instant exception and rollout details are based on OpenAI's announcement.

Hi! I frequently update with various articles about technology, practical tips, and cutting-edge news. I hope it will be helpful to you!
Last updated on 2026-11-01