ChatGPT and Codex will share usage limits: what changes in December

Short answer: Starting in December 2026, ChatGPT Plus, Pro and Business subscribers will have one usage allowance shared across ChatGPT and Codex, measured in tokens, instead of separate limits for chatting and for agent and coding work. Instant chat won't count against it, and OpenAI says it expects most people won't hit resets more often than they do now. OpenAI hasn't published how many tokens each plan includes or the exact date the change takes effect.

If you pay for ChatGPT and use Codex for coding, or you're weighing ChatGPT against Claude, Claude Code or Cursor, this changes how far your subscription goes. Below we cover what OpenAI announced, what's changing and what isn't, why some subscribers are unhappy, how it compares with rivals, and what to do before December. We've worked from OpenAI's own help pages and pricing docs, and we say so where something comes from press reports.

What OpenAI announced

OpenAI announced the change over the weekend of October 31 to November 1, 2026, in a Help Center article and an email to subscribers. The help article says that "starting in December, Plus, Pro, and Business plans' usage allowance will be shared across ChatGPT and Codex as we bring the agentic capabilities of Work into the main ChatGPT experience."

That second part is the bigger product change. According to the email, which BleepingComputer and TechCrunch reported, OpenAI will merge the separate Chat and Work tabs so you can go from a normal conversation to a longer task, such as building a website, creating a presentation or doing things on the web, without switching modes. The combined experience will be tested with a small group of users first before it rolls out more widely. Codex, OpenAI's coding agent, will stay a separate experience in the ChatGPT desktop app.

What changes, and what stays the same

From OpenAI's help article and its Codex pricing page:

• One allowance for ChatGPT and Codex. On Plus, Pro and Business, your ChatGPT use and your Codex use will draw from the same pool.
• Measured in tokens. Usage will be measured in tokens, the units of text a model reads and writes. OpenAI says more advanced models and more complex tasks generally use more of your allowance.
• Instant chat is exempt. OpenAI says "Instant chat will not draw from your usage", so quick everyday questions in Instant chat shouldn't eat into the allowance.
• One place to check. Your combined remaining usage and reset times will appear in Settings → Usage in ChatGPT.
• Same options at the limit. If you run out, you can wait for the reset, upgrade your plan, or buy credits for eligible features.

Part of this is already true. OpenAI's Codex pricing page says "ChatGPT Work and Codex share usage", and that ChatGPT for Excel on Plus and Pro shares the same limits. The New Stack reports that Codex, ChatGPT Work, ChatGPT for Excel and Workspace Agents currently draw from a shared agentic allowance, while regular ChatGPT conversations have their own limits. December's change brings everyday chat with the more capable models into the same pool.

What OpenAI hasn't said matters just as much. Neither the help article nor the email lists token amounts for any plan, and OpenAI hasn't given an exact December date. The New Stack says it asked OpenAI both questions, and whether the combined allowance would match today's separate ones, and that the company declined to answer, pointing instead to its recent limit increases.

The October 29 Pro changes

The announcement came days after a separate change to Pro. On October 29, 2026, OpenAI raised Pro allowances, and its Pro tiers page now explains the details: Pro 100 and Pro 500 now include twice their previous Work and Codex allowance, and the reduction to Pro 200 that OpenAI had announced in a September 29 email was cancelled. Pro 200 subscribers on the lower allowance had it doubled to match the higher one. The New Stack reports that Business Premium allowances were doubled too. OpenAI's page says Plus usage is unchanged.

OpenAI executive Thibault Sottiaux, who posts as Tibo, announced the reversal on X. OpenAI's page also adds a caution: a larger allowance "does not guarantee twice as many completed tasks", because usage depends on the model, task and settings.

Why some subscribers are worried

The concern is simple: once chat and Codex share a pool, a heavy coding session could leave less room for ChatGPT, and the other way round. A thread on OpenAI's developer forum asks OpenAI to keep the two allowances separate, or to guarantee subscribers won't end up with less usable capacity. The New Stack quotes the thread's author asking for usage to be broken down "by product, model, and task". On Hacker News, the discussion passed 220 points and 160 comments, BetaNews reported, with many people criticising token-based metering when no plan says how many tokens it includes.

OpenAI's answer so far is the line in its help article: "We expect most users won't need to wait for their usage to reset more often than they do now." TechCrunch's reading of the email is that, depending on usage and model, some people may need to wait for resets more often. Until OpenAI publishes per-plan numbers, it's hard to say who will come out ahead.

How this compares with Claude and Cursor

Shared pools are becoming the norm for AI subscriptions that include a coding agent.

• Claude. Anthropic's help center says Pro, Max and Ultra limits are "shared across Claude and Claude Code, meaning all activity in both tools counts against the same limits". Pro and Max have a five-hour session limit and a weekly limit, and Max and Ultra also have a separate weekly limit for Fable. Anthropic also merged Claude's chat and Cowork into one screen in September 2026, the same kind of merge OpenAI is now testing.
• Cursor. Cursor's pricing docs describe one weekly pool for Cursor's own models, shared across Agent in the editor, the CLI, Cloud Agents and the Grok apps. Third-party models such as Claude and GPT are billed separately from prepaid credit.
• ChatGPT and Codex. From December, the same idea applies to OpenAI's plans. The difference for now is visibility: OpenAI publishes rough Codex message estimates, but not the size of the new combined token allowance.

So the real question isn't whether limits are shared. It's how big the pool is, how quickly your usual models drain it, and how clearly each tool shows you what's left. Our Claude vs ChatGPT, Claude Code vs Cursor and OpenClaw vs Claude Code comparisons cover the wider trade-offs.

What to do before December

• Check your current usage. Look at Settings → Usage in ChatGPT, or Usage & Billing in Codex, to get a feel for how often you hit limits today.
• Use Instant for quick questions. Since Instant chat won't count against the shared allowance, save the heavier models for work that needs them.
• Watch speed modes. OpenAI says Fast mode uses your included allowance at 2.5 times the Standard rate, and Ultrafast (on Pro 500) at 8 times. These will matter more once chat and Codex share a pool.
• Know your fallbacks. OpenAI says Plus and Pro users can buy credits after reaching the limit, and anyone can run extra local Codex chats with an API key at standard API rates. Check the live pricing page for current rates.
• Re-check in December. If OpenAI publishes per-plan allowances, compare them with what you use now. If you mainly use ChatGPT for chat and something else for coding, or the reverse, the merged pool may affect you differently from someone who uses both heavily.

What it means if you're choosing AI tools

For most people, this is a billing change rather than a new feature, and OpenAI says most users won't notice. It matters most if you use ChatGPT heavily for research or writing and run long Codex sessions on the same plan. If that's you, it's worth tracking usage now so you can tell in December whether your plan still fits, and comparing what Claude or Cursor offer at a similar tier. For the latest on OpenAI's models, see our GPT-6.1 Terra explainer and what to expect from GPT-7.

FAQ

When do ChatGPT and Codex start sharing usage limits?

OpenAI says the shared allowance starts "in December" 2026 for Plus, Pro and Business plans. It hasn't given an exact date. The Chat and Work tab merge will be tested with a small group of users first.

Will the shared allowance mean less usage for me?

OpenAI says it expects most users won't need to wait for resets more often than they do now. It hasn't published per-plan token amounts, and it declined to tell The New Stack whether the combined allowance would match today's separate ones, so you can't check this for yourself yet.

Does normal ChatGPT chat count against the shared limit?

Instant chat won't draw from the allowance, according to OpenAI's help article. Using more advanced models and complex tasks will draw from it, and these generally use more of your allowance.

What happens if I hit the ChatGPT and Codex usage limit?

You can wait for the reset, upgrade your plan, or buy credits for eligible features such as Codex, ChatGPT Work and ChatGPT for Excel. Your combined remaining usage and reset times will show in Settings → Usage.

Do Claude and Claude Code share usage limits too?

Yes. Anthropic says Pro, Max and Ultra usage limits are shared across Claude and Claude Code, so activity in both counts against the same limits. Pro and Max have five-hour and weekly limits.