ChatGPT and Codex Will Share One Usage Limit From December 2026: What It Means for You

The ChatGPT and Codex shared usage limit is coming in December 2026, and it will change how you plan your AI usage. If you pay for ChatGPT Plus, Pro or Business and also use Codex for coding, there’s a change coming in December 2026. OpenAI has updated its Help Center to say that ChatGPT and Codex will share one usage allowance from December, so you won’t have two separate quotas any more.

That might sound like a small billing detail, but it changes how students, developers and working professionals should plan their AI usage every day. Let’s break it down in simple words.

What exactly is changing in the ChatGPT and Codex shared usage limit?

Right now, ChatGPT (the chat app) and Codex (OpenAI’s AI coding agent) have separate usage limits. Using Codex heavily does not cut into your ChatGPT chats, and the other way round.

From December, according to OpenAI’s Help Center article on ChatGPT and Codex usage limits, eligible plans will have one combined pool. Think of it like a family mobile data plan: everyone takes from the same bucket. If you use a lot in Codex, you have less left for advanced ChatGPT features before the reset, and vice versa.

Here are the main points reported so far:

• Who it affects: ChatGPT Plus, Pro and Business plans. The reports we checked do not mention the Free or Go plans.
• When: December 2026. OpenAI has not given an exact date yet.
• How it is measured: in tokens, not simply the number of messages.
• What does not count: Instant chat reportedly does not draw from the shared allowance.
• Where to check: your combined remaining usage and reset times will appear under Settings → Usage.
• At the limit: you can wait for the reset, upgrade your plan, or buy credits to continue.

ChatGPT and Codex shared usage limit: first, what are tokens?

A token is a small piece of text that an AI model reads or writes. Roughly, one token is about three-quarters of an English word, so a long answer uses more tokens than a short one.

Because the new system counts tokens, not every request costs the same. Two things matter:

• The model you pick: more advanced reasoning models generally use more of your allowance.
• The size of the task: asking Codex to fix one function is cheap; asking it to refactor a whole project, run tests and repeat is much more expensive.

So a single big Codex job could use as much allowance as many normal chat questions.

Why is OpenAI creating a ChatGPT and Codex shared usage limit?

OpenAI has not published a long explanation, so we should be careful not to guess too much. But the idea of one pool is easy to understand:

• Simpler for users: one number to track instead of two separate limits.
• Flexible: someone who mostly chats can use most of the pool in ChatGPT, while a developer can spend most of it in Codex.
• Fairer cost tracking: AI agents like Codex can run long tasks that cost a lot to compute. Counting tokens connects your usage to the real work done.

OpenAI has reportedly said it expects most users will not need to wait for resets more often than they do today. We will only know if that is true after the change goes live.

What is still not clear about the ChatGPT and Codex shared usage limit?

Several details are still missing, so treat any exact numbers you see online with caution:

• The exact start date in December.
• How many tokens each plan (Plus, Pro, Business) will get.
• How the limits are split between short windows and weekly windows. Reports suggest there may be a 5-hour limit and a weekly limit, and both must have room left for you to continue.
• Which ChatGPT features, such as image generation, will take from the shared pool.

One more point: a separate OpenAI page about using Codex with a ChatGPT plan reportedly still showed the older limits at the time of writing. So OpenAI’s own pages were not fully aligned yet. Always check the latest official Help Center page before making decisions. Related reading: Texas sues OpenAI over ChatGPT safety.

Who should care about the ChatGPT and Codex shared usage limit the most?

1. Students and freshers learning to code

If you use ChatGPT to understand concepts and Codex to practise coding projects, both will now come from the same budget. Long Codex sessions before an exam or a hackathon could leave you with less chat usage that day.

2. Working developers

Developers who let Codex run long agent tasks in the background will feel this most. Large refactors, test loops and code reviews can use a lot of tokens quickly.

3. Teams on Business plans

Team leads may want to explain the change to their members, so nobody is surprised when they hit a limit in the middle of a deadline.

4. Light users

If you mostly ask quick questions, you may hardly notice any difference, especially if Instant chat does not count toward the pool as reported.

What should you do about the ChatGPT and Codex shared usage limit?

• Check Settings → Usage regularly once the change is live, so you know how much is left and when it resets.
• Pick the right model for the job. Don’t use the heaviest reasoning model for a simple grammar check or a basic definition.
• Break big Codex tasks into smaller steps. Clear, focused instructions usually waste fewer tokens than vague “fix everything” prompts.
• Plan heavy work. If you have a big coding job, avoid also doing hours of advanced chat in the same window.
• Know your options at the limit: wait for the reset, upgrade, or buy credits. Compare the cost before buying credits regularly.
• Follow official updates. Rely on OpenAI’s Help Center rather than screenshots shared on social media.

Key takeaways on the ChatGPT and Codex shared usage limit

• From December 2026, ChatGPT and Codex will share one usage allowance on Plus, Pro and Business plans.
• Usage is counted in tokens, so advanced models and big tasks use more.
• Instant chat reportedly does not count toward the shared pool.
• Exact per-plan amounts and the start date are not public yet.
• Smart model choice and smaller tasks will help your allowance last longer.

FAQ

Sources