From December 2026, ChatGPT and Codex will share one usage allowance on Plus, Pro and Business plans. OpenAI will measure that allowance in tokens, so heavier models and longer tasks use it up faster, while quick Instant chats won’t count at all. OpenAI hasn’t named the exact December date and hasn’t said how many tokens each plan gets. Below we set out what OpenAI has confirmed, how the limits work today, and which plan fits which kind of use. We read OpenAI’s Help Center pages and the subscriber email reported by BleepingComputer on 3 November 2026. We haven’t measured any limits ourselves, so the numbers here are OpenAI’s own.
- One allowance from December. On Plus, Pro and Business, ChatGPT and Codex will draw from the same pool. No exact start date yet.
- Counted in tokens, not messages. The model, the task and your settings decide how fast you use it. Instant chat doesn’t count.
- Work and Codex already share a pool today. Plus and Business Standard have a five-hour window and a weekly window. Pro 100, Pro 200 and Pro 500 currently have no five-hour limit there.
- Pro allowances went up on 29 October. Pro 100 and Pro 500 doubled, Pro 200’s planned cut was cancelled, and Plus stayed the same.
- Your balance lives in Settings → Usage. When it runs out you can wait for the reset, upgrade or buy credits. Switching models doesn’t give you allowance back.
What changes in December?
OpenAI’s Help Center article “Understanding ChatGPT and Codex usage limits”, updated around 1 November 2026, says: “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.” In practice that means five things.
- One allowance for two products. What you use in ChatGPT and what you use in Codex comes out of the same pool.
- Tokens decide the cost. Tokens are the units of text a model reads and writes. More advanced models and more complex tasks use more of them.
- Instant chat is left out. OpenAI says Instant chat “will not draw from your usage”. The subscriber email calls it “Unlimited Instant”.
- One place to check. Settings → Usage shows your combined remaining usage and when it resets.
- Three ways forward at the limit: wait for the reset, upgrade your plan, or buy credits for eligible features.
OpenAI also says “We expect most users won’t need to wait for their usage to reset more often than they do now.” That is an expectation, not a promise, and without per-plan token figures nobody outside OpenAI can check it yet.
Work features are moving into Chat
The shared allowance comes with a product change. In an email to subscribers that BleepingComputer reported on 1 November 2026, OpenAI said it “will test a combined experience for Chat and Work tabs in the coming weeks, initially with a small group of users”. In that combined view you could go from an ordinary question to a longer task without switching tabs. OpenAI’s own example: “You can bring in your files and plugins and ask ChatGPT to build a website, create a presentation, or do things for you on the web, while reviewing and guiding its work.”
Codex stays where it is. The same email says Codex will remain a separate experience in the ChatGPT desktop app; only the allowance behind it is shared. If you are still deciding between coding agents, our Codex vs Claude Code comparison covers how the two work day to day.
How the limits work today
Part of the pool is already shared. According to OpenAI’s page on managing usage with GPT-6 Astra, Work and Codex draw from the same included allowance on Plus and Pro and on Business Standard and Premium seats. On plans with shared allowances or credits, ChatGPT for Excel and Workspace Agents use that pool too.
- It isn’t a fixed number of messages. How much you get done depends on the model, the task and your settings. A long agentic task can use far more than a short request.
- Two windows can apply. Depending on the plan there is a five-hour window and a weekly window, and you need allowance left in both. A new five-hour window starts with your first Work or Codex message after the previous one ends, so you can reach the five-hour limit well before five hours have passed.
- Pro has no five-hour limit for now. Pro 100, Pro 200 and Pro 500 currently have no five-hour limit in Work and Codex. They still have an included allowance.
- Some settings cost more. Larger inputs and outputs, higher reasoning effort, Fast mode and multi-step tasks all raise usage. Fast mode uses included allowance at 2.5 times the Standard rate.
For Plus and Business Standard, OpenAI publishes estimated ranges of local messages per five-hour window in Work and Codex. The figures are the same for both plans.
| Model (as named in OpenAI’s estimate table) | Local messages per five-hour window, Plus and Business Standard |
|---|---|
| GPT-6 Astra | 5–45 |
| GPT-6.1 Sol | 15–160 |
| GPT-6 Sol | 15–150 |
| GPT-6 Luna | 350–3,000 |
OpenAI stresses that these are estimates, not guaranteed counts. Cloud tasks may use more than local messages, weekly limits may also apply, and API usage is billed separately. The ranges are wide because one message can be a one-line fix or an hour-long task. OpenAI also says that “Astra at Low effort can outperform Sol at High effort”, so the cheapest-looking model isn’t automatically the most economical one for hard work.
Some chat limits sit outside the pool today. In Chat, GPT-6 Pro is powered by Astra and is available on eligible Pro, Business and Enterprise plans, with message limits that OpenAI says are separate from the Work and Codex allowance. Plus includes Astra in Work and Codex but not GPT-6 Pro in Chat. Regular Chat Voice, file uploads and image generation also have their own limits and don’t consume Codex usage. OpenAI hasn’t said whether any of this changes in December.
Pro tiers and the 29 October changes
OpenAI’s “About ChatGPT Pro tiers” page lists three Pro plans. All are billed monthly; there is no annual option. Prices are in US dollars, and local prices and taxes may differ.
| Plan | Per month | Work and Codex allowance since 29 Oct | Ultrafast |
|---|---|---|---|
| Pro 100 | $100 | Doubled | Not included |
| Pro 200 | $200 | Planned cut cancelled; lower allowance doubled to match the higher one | Not included |
| Pro 500 | $500 | Doubled | Included |
The Pro 200 story needs a word of context. A reduction of its allowance had been announced in an email on 29 September. OpenAI has now cancelled it: subscribers on the lower allowance had it doubled to match the higher one, the price stays at $200, and Pro 200 is open to new subscribers again. Weekly GPT-6 Pro message limits in Chat stay the same on Pro 100 and Pro 500, and Pro 200’s lower limit was raised to match the higher one. Plus usage is unchanged. OpenAI adds that “A larger allowance does not guarantee twice as many completed tasks.”
Ultrafast is the only feature reserved for one self-serve plan. On Pro 500 it is available for GPT-6 Astra and GPT-6.1 Sol and uses included usage at 8 times the Standard rate, then credits at 6 times the Standard credit rate. Plus, Pro 100, Pro 200 and Business don’t get it, and buying credits doesn’t unlock it.
Which plan fits your usage
We built this table only from what OpenAI publishes. It doesn’t rank the plans by how much work they get through, because OpenAI doesn’t publish allowance sizes in tokens. Treat it as a way to rule plans in or out by the limits and features that apply to you.
| Plan | Five-hour window in Work and Codex | GPT-6 Pro in Chat | Ultrafast | Fits you if |
|---|---|---|---|---|
| Plus | Yes plus weekly window | No | No | You mostly chat, use Instant for quick questions and run Codex or Work in short sessions. You can live with pauses when the five-hour window runs out. |
| Pro 100 | None for now | Yes, eligible accounts | No | You hit the Plus five-hour window regularly and want long agentic or coding sessions without that stop, or you need GPT-6 Pro in Chat. |
| Pro 200 | None for now | Yes, eligible accounts | No | You were already on it, or Pro 100 runs dry before the week is out. OpenAI doesn’t publish how its allowance compares with Pro 100 in tokens, so watch Settings → Usage before moving up. |
| Pro 500 | None for now | Yes, eligible accounts | Yes | You need Ultrafast for GPT-6 Astra or GPT-6.1 Sol and accept that it burns included usage at 8 times the Standard rate. |
| Business (Standard seat) | Yes same estimates as Plus | Yes, eligible accounts | No | You need a workspace for a team rather than a personal plan, and per-person usage like Plus is enough. Premium seats also share Work and Codex. |
Two rules of thumb follow from OpenAI’s pages. If the five-hour window is what stops you, any Pro tier removes it for now. If the weekly window is what stops you, a bigger plan helps, but so does choosing a lighter model or a lower effort setting for routine work. If you are weighing ChatGPT against other paid assistants rather than against its own tiers, our guide to Claude, ChatGPT or Gemini compares the subscriptions.
What counts against the shared allowance?
OpenAI hasn’t published a complete list for December. This is what its pages say so far.
| Activity | Counts against the shared allowance? | Status |
|---|---|---|
| Instant chat | No | Confirmed for December |
| Codex, local and cloud tasks | Yes | Confirmed |
| ChatGPT Work and agentic tasks | Yes | Confirmed |
| ChatGPT for Excel, Workspace Agents | Yes, on plans with shared allowances or credits | Confirmed today |
| Other ChatGPT chats with advanced models | Expected to, measured in tokens; detail not published | Not yet detailed |
| GPT-6 Pro in Chat | Separate message limits today | December not stated |
| Regular Chat Voice | Separate limits today; tasks started by voice in Work or Codex on desktop do count | December not stated |
| File uploads, image generation | Separate limits today, don’t apply to Codex | December not stated |
| Codex auto-review checks | No, they are free | Confirmed today |
Check your usage in 30 seconds
This checklist follows the steps in OpenAI’s help pages. Menu names may shift as the combined Chat and Work view rolls out.
- Open Settings → Usage in ChatGPT. It shows your remaining usage and reset times; from December it shows the combined figure for ChatGPT and Codex.
- Note which window is low. On Plus and Business Standard, check the five-hour and the weekly window separately. You need both.
- In a Codex CLI session, type
/statusto see your Codex usage without leaving the terminal. - Look for a banked reset. Saved resets appear in Settings → Usage in the desktop app and can be used there.
- Check what you are running. Model, reasoning effort and Fast mode all change how fast the balance drops. Fast mode costs 2.5 times the Standard rate.
What to do if you hit the limit
- Wait for the reset. The time is shown in Settings → Usage. If you hit the limit in the middle of a Codex turn, Codex can finish that turn.
- Use a banked reset if you have one. Banked resets, for example from a referral, refresh both the five-hour and the weekly window and move your weekly reset date.
- Buy an instant reset. Eligible personal Plus and Pro accounts can buy one, depending on billing country. It refreshes both windows straight away, can’t be saved for later and moves your weekly period.
- Buy credits. Included usage is used first. After that, credits cover eligible usage in Codex, ChatGPT Work and ChatGPT for Word, Excel and PowerPoint.
- Upgrade to a plan with a larger allowance or no five-hour window.
Two things won’t help. Switching models doesn’t restore allowance in a shared pool, and OpenAI says support can’t reset limits for you. If you keep running into the ceiling and want a second assistant for overflow work, our list of ChatGPT alternatives covers what each one does better.
+What the change gets right
- One balance and one usage page instead of separate meters
- Instant chat stays outside the limit
- Pro allowances were raised, not cut, before the switch
- Several ways to keep working at the limit: banked resets, instant resets, credits
−What is still unclear
- No token figures per plan, so plans can’t be compared exactly
- No exact December start date
- Heavy chat could now eat into Codex time, and the other way round
- Model names differ between OpenAI’s own lists
What we don’t know yet
- The exact date in December when the shared allowance starts.
- How many tokens each plan includes.
- Whether the five-hour and weekly windows stay as they are once ChatGPT joins the pool, and whether Pro keeps having no five-hour limit.
- How Free, Go and Enterprise are affected. The December notice names only Plus, Pro and Business, although Codex itself is included on Free and Go.
- When the combined Chat and Work view reaches everyone.
- Which model names in OpenAI’s estimate table are current.
For most people on Plus the change should feel small: OpenAI expects reset waits to stay about where they are, and Instant chat, which covers a lot of everyday questions, stays free of the meter. The people to watch are those who use ChatGPT heavily and Codex heavily on the same plan, because from December both will draw on one balance. Before you upgrade, spend a week looking at Settings → Usage and see which window actually stops you. If it is the five-hour window, Pro 100 removes it for now. If it is the weekly one, try a lighter model or lower effort for routine work first. For a broader view of how ChatGPT compares with its closest rival, see Claude vs ChatGPT.
Frequently asked questions
When do the new ChatGPT usage limits start?
OpenAI says “starting in December” 2026. It hasn’t given an exact date.
Which plans get the shared ChatGPT and Codex allowance?
Plus, Pro and Business, according to OpenAI’s Help Center. Free, Go and Enterprise aren’t named in the notice.
Does Instant chat count towards the limit?
No. OpenAI says Instant chat will not draw from your usage, and its email to subscribers calls it unlimited.
Does ChatGPT Pro have a five-hour limit?
Not in Work and Codex at the moment. Pro 100, Pro 200 and Pro 500 currently have no five-hour limit there, but they still have an included allowance.
Where can I see my remaining usage?
In Settings → Usage in ChatGPT, which shows remaining usage and reset times. In a Codex CLI session you can type /status.
Does switching to a smaller model give me more allowance?
It can make the allowance last longer, because lighter models use fewer tokens. It doesn’t restore allowance you have already used in a shared pool.
Sources and method
Limits, windows, model estimates, Pro tiers, prices, resets and credits are taken from OpenAI’s Help Center articles, checked on 3 November 2026. Plan prices are in US dollars. The product plans for Chat and Work come from OpenAI’s subscriber email as reported by BleepingComputer; where BleepingComputer and OpenAI word the Pro changes differently, we follow OpenAI. The plan table and the usage checklist are our own summary of these pages. We haven’t measured usage limits on our own accounts.
- OpenAI Help Center, Understanding ChatGPT and Codex usage limits (updated around 1 November 2026)
- OpenAI Help Center, Managing usage with GPT-6 Astra in Work and Codex (updated 2 November 2026)
- OpenAI Help Center, About ChatGPT Pro tiers (updated 2–3 November 2026)
- OpenAI Help Center, Using Codex with your ChatGPT plan (updated 2–3 November 2026)
- BleepingComputer, OpenAI says it’s doubling Pro usage limits and bringing Work features to Chat (1 November 2026)
