On October 29, 2026, OpenAI did something its paying customers had stopped expecting: it gave them more for the same money, and it cancelled a price-hike-in-disguise the day before it was due to hit. Thibault Sottiaux — “Tibo” on X, the OpenAI lead for Codex and ChatGPT Work — posted that the company was “doubling all usage limits for all Pro 100/200/500 tiers and Business Premium, effective today,” and that the planned Pro 200 cut was off: “We heard your kind feedback and won’t be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same.”
If you skimmed the headlines, you probably came away thinking everyone on a Pro plan just got twice as much. That is not quite what happened. The real story is better than the headline for some subscribers, a non-event for others, and a genuine reprieve for one group that was about to lose half its allowance overnight. This article sorts out who actually gained what, what did not change, and the three checks to run on your own account before you upgrade, downgrade, or renew.
What actually changed on October 29
The change covers ChatGPT Work and Codex, which share a usage allowance under eligible subscriptions. It does not cover everything with the word “limit” on it — that distinction is the whole story, so let us be precise.
Doubled, effective immediately, at unchanged prices: the Work and Codex allowances on Pro 100 ($100/month), Pro 500 ($500/month), and Business Premium. No action required; existing subscribers got the higher allowance automatically, without waiting for renewal.
The planned Pro 200 cut, cancelled: OpenAI had announced on September 29 that existing Pro 200 ($200/month) subscribers would drop from 20x the ChatGPT Plus allowance to 10x, and that weekly GPT-6 Pro chat messages would fall from 200 to 100. The cut was due October 30. It is now off the table entirely. Subscribers who were on the higher allowance keep it; subscribers who had been put on the lower allowance when the plan reopened were raised to the higher one.
Dots, untouched and separate: OpenAI’s always-on assistants, called dots, keep their extended limits, and dots usage stays separate from the Work and Codex allowance. (We covered what dots actually do for a small team separately.) Sottiaux’s wording — extended limits continue “for now” — is worth noting: it is a promise with a visible expiry of confidence, not a permanent guarantee.
What did not change: ChatGPT Plus limits are exactly what they were. The weekly GPT-6 Pro message limits in regular Chat on Pro 100 and Pro 500 stay where they were. Ultrafast access remains exclusive to Pro 500. And no subscription price moved in either direction.
One more honesty note before the tables: OpenAI does not publish exact numerical usage limits per tier. The doubling is real as an announced policy, but no subscriber can verify it against a published number — you can only watch your own usage meter and see whether you hit the wall later than you used to. Every article that gives you a precise “you now get X messages” figure for this change is inventing it. This one will not.
The tier truth table: who gained, who merely survived, who got nothing
| Plan | Price | What actually changed Oct 29 |
|---|---|---|
| ChatGPT Plus | $20/month (unchanged) | Nothing. No extra allowance. |
| Pro 100 | $100/month | Work and Codex allowance doubled. Weekly GPT-6 Pro chat cap unchanged. |
| Pro 200 (existing, higher allowance) | $200/month | Nothing gained — but a 50% cut avoided. You keep the 20x allowance you were about to lose. |
| Pro 200 (newer, lower allowance) | $200/month | Raised to the higher allowance — effectively doubled back to the level existing subscribers already had. |
| Pro 500 | $500/month | Work and Codex allowance doubled. Ultrafast access unchanged (still exclusive to this tier). |
| Business Premium | Per-seat (unchanged) | Work and Codex allowance doubled. OpenAI gives no per-seat figure. |
| Dots (all eligible plans) | Included | Extended limits continue, metered separately. Described as continuing “for now.” |
Read that table again if you are on Pro 200, because the coverage got this wrong in both directions. Some headlines said Pro 200 users’ limits doubled. Others said nothing changed for them. Both are half true, depending on which Pro 200 subscriber you are. If you held the plan before the September 29 announcement and kept the 20x allowance through the grace period, your allowance today is the same as it was on October 28 — the win is the cut that did not happen, not a doubling. If you joined after the plan reopened on the lower allowance, you genuinely got more. The distinction matters most if you were about to switch plans to escape the cut: that switch may now be unnecessary, and it is worth re-checking before you pay for a different tier.
There is also a reported loose end worth flagging honestly. When the cut was announced in September, reporting based on user-shared emails described a one-time grant of 62,500 usage credits, valued by OpenAI at $2,500 and expiring December 31, 2026, as compensation for existing Pro 200 subscribers. OpenAI has not said publicly what happens to that grant now that the cut is cancelled. Do not count on it, and do not assume it vanished either — check your account’s credit balance rather than trusting either assumption.
Check your own allowance in 3 steps
Because OpenAI publishes no numbers, the only reliable test is your own account. Here is the short version.
Step 1: Confirm which Pro tier you are actually on. In ChatGPT, open Settings → Subscription (or your plan page) and confirm whether you hold Pro 100, Pro 200, or Pro 500, and whether your plan is a legacy grandfathered variant. The tier name determines which row of the truth table applies to you. OpenAI’s Pro tiers help page lists what each tier includes. If you switched plans in late September or October to dodge the announced cut, this is the step that tells you whether to switch back.
Step 2: Look at your usage meter, not your plan page. The plan page will show the tier name and price; it will not show you a numerical allowance. Your usage meter — in ChatGPT Work or Codex, where you hit rate limits in practice — is the only place the change is observable. If you regularly hit your cap mid-week on Pro 100, the test is simple: does the cap now arrive noticeably later, or not at all, on the same workload? Give it a full billing cycle before concluding anything, since a quiet week proves nothing.
Step 3: Decide whether the doubling changes your plan. This is where the money is. If you moved from Pro 100 to Pro 200 in October purely because you expected the Pro 100 allowance to stay small, the doubled Pro 100 allowance may now cover your work — a $100/month saving. Conversely, if you are on Pro 200 and were considering Pro 500 for the headroom, the preserved 20x allowance plus the doubling elsewhere in the lineup may mean you no longer need the jump. And if you are on Plus and wondering whether this news is a reason to upgrade: it is not. Plus got nothing from this announcement, and the upgrade decision should rest on whether you need Work and Codex at all, which brings us to the business question.
The small-business decision: doubled Pro vs pay-as-you-go API
For a solo operator or a small team, the interesting question is not “did limits go up” but “should I be paying a subscription at all?” The October 29 change shifts that math in one specific direction: it makes Pro 100 a stronger deal for people whose usage is bursty but real — the freelancer who runs Codex hard for two weeks on a client project, then barely touches it for a month.
Think of it this way. Pro 100 at $100/month buys you a doubled Work and Codex allowance with no metering anxiety during the month. The API route, by contrast, bills per token: cheap in a quiet month, uncapped in a heavy one, and it requires you to manage keys, rate limits, and cost alerts yourself. The subscription wins when your monthly usage is lumpy and your time spent watching meters is worth something. The API wins when your usage is steady and predictable, or when you only need frontier models occasionally — especially after the October 27 launch of GPT-6.1 Terra at half the price of GPT-5.6 Terra, which lowered the API side of the comparison in the same week.
A practical rule of thumb: if you hit your Pro 100 cap at least twice in the last two months, the doubling probably just saved you an upgrade to Pro 200 — stay put and re-measure. If you have never hit the cap, the doubling is worth exactly nothing to you, and your decision should be about features (Ultrafast on Pro 500, dots, Business Premium seats) rather than allowances. For the wider OpenAI picture, see what OpenAI’s always-on dots mean for small business and the GPT-6 launch coverage. The tier-by-tier reporting behind this article is documented in the ghacks breakdown and the plan-by-plan check. And if you are a team, Business Premium’s doubled allowance matters more than any individual tier change, because shared seats burn through a pooled allowance faster than any single user — but OpenAI gives no per-seat figure, so verify with your actual team’s usage before adding seats on the strength of this announcement.
Why OpenAI reversed course (and why it matters for your planning)
The reversal did not happen in a vacuum. The September 29 cut announcement landed badly: Pro 200 subscribers who had paid $200 a month for a 20x allowance were told it would halve to 10x at the same price, with a one-time credit grant as compensation. The reaction on X was loud — one widely shared post called the grant a one-off payment that “doesn’t make up for a permanent cut” — and the reversal post itself drew hundreds of thousands of views within hours. OpenAI framed the reversal as listening to feedback, and whatever the internal calculus, the sequence teaches a planning lesson: Pro-tier allowances are now visibly negotiable, and they can move in both directions on a month’s notice.
That cuts both ways for subscribers. The doubling is good news today, but it was announced in a post on X, not in a contract amendment, and the dots limits carry an explicit “for now.” If your business depends on a specific allowance — if you have client deadlines riding on Codex capacity — treat the current limits as the current weather, not the climate. Keep an API fallback configured, keep your prompts and workflows portable, and do not sign a client commitment that assumes today’s allowance will still exist in six months. The same caution applies to the broader AI vendor landscape: a vendor that can halve your allowance with a month’s notice can also double it, and neither move should be the foundation of your capacity planning.
The September timeline, because the order matters
The October 29 reversal only makes sense against the six weeks that preceded it. On September 29, Sottiaux announced that the reopened Pro 200 plan would carry a lower allowance — 10x Plus instead of 20x — and that existing subscribers would move to the lower allowance after a grace period ending October 29, with a one-time credit grant as compensation. On September 30, the plan reopened to new subscribers at that lower allowance. For a month, two classes of Pro 200 subscriber existed side by side at the same $200 price: legacy holders on 20x through the grace period, and new joiners on 10x from day one. The October 29 announcement collapsed that two-tier arrangement back into one: the lower tier was raised, the grace period became moot, and the cut scheduled for October 30 never happened. If you joined Pro 200 in October and felt shortchanged next to a colleague paying the same price, that feeling was accurate — and it is now resolved in your favor. If you are still unsure which class you were in, your October billing history and the plan-change notices in your email will tell you; OpenAI’s own help page confirms the cut’s cancellation but does not reconstruct your personal timeline for you.
What to watch next
Three things will tell you whether October 29 was a durable improvement or a tactical retreat. First, whether OpenAI publishes actual numbers: the absence of per-tier figures is the single biggest gap in this story, and if the company ever discloses them, every subscriber can finally verify what they bought. Second, whether the dots limits survive their “for now” — dots are the product line OpenAI is pushing hardest for always-on work, and their separate metering is the most likely place for the next pricing change. Third, whether Pro 500’s Ultrafast access stays exclusive or trickles down: exclusivity is Pro 500’s main remaining differentiator now that allowances have doubled below it, and OpenAI has a habit of moving premium features downmarket once they are no longer scarce.
For now, the practical summary is short. Pro 100 and Pro 500 subscribers: you got twice the Work and Codex room at the same price. Pro 200 subscribers: you kept what you had, and if you were on the lower allowance, you got raised. Plus subscribers: nothing changed, and this announcement is not a reason to upgrade. Business Premium teams: your pooled allowance doubled, but verify it against your actual usage. And everyone: the only limit figure you can trust is the one your own usage meter shows you over the next billing cycle.
Frequently Asked Questions
Did OpenAI really double limits for all Pro plans?
For Work and Codex usage, yes — Pro 100, Pro 200, and Pro 500, plus Business Premium, effective October 29, 2026. But the details differ by tier: Pro 200 subscribers already on the higher allowance saw no increase, only the cancellation of a planned cut. Weekly GPT-6 Pro chat message limits and Plus limits did not change.
What was the cancelled Pro 200 cut?
Announced September 29, 2026, it would have reduced existing Pro 200 subscribers’ Work and Codex allowance from 20x the Plus allowance to 10x, and cut weekly GPT-6 Pro chat messages from 200 to 100, starting October 30. After subscriber backlash, OpenAI cancelled it on October 29, the day before it was due.
Do I need to do anything to get the doubled limits?
No. OpenAI applied the higher allowances automatically to eligible accounts. There is no opt-in and no need to wait for renewal. If you switched plans to dodge the cancelled cut, though, it is worth reviewing whether that switch is still needed.
How can I verify my new limit if OpenAI publishes no numbers?
You cannot verify it against a published figure, because none exists. The practical test is your own usage meter in ChatGPT Work or Codex: if you previously hit your cap on a known workload, see whether the cap now arrives later or not at all over a full billing cycle.
Did ChatGPT Plus get anything from this announcement?
No. Plus limits are unchanged. The doubling covers Pro 100/200/500 and Business Premium only, and only for Work and Codex usage.
Are dots usage limits included in the doubling?
No. Dots — OpenAI’s always-on assistants — keep separate extended limits that do not draw from the Work and Codex allowance. OpenAI describes those extended limits as continuing “for now,” so treat them as the least permanent part of the announcement.