OpenAI has reversed its planned ChatGPT Pro 200 usage cut. It is also doubling the Work and Codex allowance on Pro 100 and Pro 500, while bringing lower-allowance Pro 200 subscriptions up to the higher level.
The distinction matters: an existing Pro 200 customer who already had the higher allowance keeps it. That customer is not being promised another doubling on top. OpenAI’s updated Pro subscription guide confirms the change on October 29, 2026.
Published October 29, 2026. Prices below are US-dollar list prices; local currency, applicable taxes and checkout terms can affect the amount charged.
For people who build with Codex or delegate substantial work to ChatGPT, the practical news is welcome: the expected reduction has been canceled, and several subscriptions now provide more capacity. The subscription price and the usage allowance are separate parts of the story.
What OpenAI changed—and who gets more usage
| Subscription or account situation | Monthly Pro price | Work and Codex change |
|---|---|---|
| Pro 100 | $100 | Twice the previous allowance |
| Pro 200 with the lower allowance | $200 | Doubled to match the higher allowance |
| Pro 200 already on the higher allowance | $200 | Existing higher allowance retained; planned reduction canceled |
| Pro 500 | $500 | Twice the previous allowance |
New subscriptions receive the updated allowance. Existing subscribers receive it automatically, without waiting for renewal. Plus usage is unchanged. Source: OpenAI’s updated Pro guide.
This is why the phrase “OpenAI doubled Pro” needs context. Two subscribers paying $200 could have been on different allowances before this announcement. The change removes that distinction rather than creating a further reward exclusively for the previously protected group.
For a customer who had been preparing for a smaller allowance tomorrow, keeping the existing capacity may be the biggest benefit. For a customer already on the lower allowance, the improvement is an actual increase from the level they were receiving.
Why the canceled Pro 200 cut matters
In our September coverage of the Pro 200 changes, the proposed Work and Codex allowance moved from the historically described 20× Plus level to 10×. Eligible existing customers were temporarily protected through October 29, with the reduction expected afterward. Those figures describe the earlier announcement, not a newly published fixed message quota.
That created an awkward purchasing decision. A professional could budget for the same monthly payment while expecting substantially less included capacity. A developer in the middle of a project also had to decide whether to upgrade, buy additional usage or change tools.
Today’s reversal changes that calculation. Someone who was considering an upgrade solely to avoid the announced reduction should reassess the subscription they now have before paying more.
Our view: retaining predictable access is particularly valuable for ongoing work. A useful agent becomes part of a routine—researching a brief, maintaining a website or helping finish a coding project. Changing the amount of work included in that routine can matter as much as introducing a better model.
ChatGPT plans and pricing: the current individual lineup
| Individual plan | Listed monthly price | Where it sits |
|---|---|---|
| Free | $0 | Entry-level access with limited advanced usage |
| Go | $8 | Lower-cost individual option |
| Plus | $20 | Paid access for broader work and productivity |
| Pro 100 | $100 | Lowest-priced Pro subscription |
| Pro 200 | $200 | Middle Pro subscription |
| Pro 500 | $500 | Highest Pro subscription; includes Ultrafast access |
The prices are listed in OpenAI’s Work and Codex pricing documentation. The ChatGPT plan comparison describes the wider bundle of chat, research, image and productivity features. This announcement concerns included usage; it is not a doubling of monthly subscription prices.
For budgeting, twelve monthly payments total $240 for Plus, $1,200 for Pro 100, $2,400 for Pro 200 and $6,000 for Pro 500, before applicable taxes or other checkout differences. These are arithmetic comparisons of monthly charges, not advertised annual subscription discounts.
The step from Plus to Pro 100 adds $80 a month. Pro 100 to Pro 200 adds another $100. Pro 200 to Pro 500 adds $300. Those differences are easier to evaluate against a real workload than the plan names alone.
A higher tier can be worthwhile when a limit repeatedly interrupts valuable work. Paying for unused headroom is harder to justify. Start with how often you run out, which activity causes it and what the interruption costs you.
Business Premium is part of the announcement too
Official-source update, October 29: OpenAI’s October 29 release notes name Pro and Business Premium in the announcement heading and say their Work/Codex usage is increasing. The inspected text specifies doubling for Pro 100, Pro 500 and lower-allowance Pro 200 accounts, but does not establish a Business Premium multiplier. We therefore do not assign Premium a new numerical ratio. Business Premium, Business Standard and personal Pro subscriptions remain distinct options.
| Business seat | Monthly billing | Annual billing: monthly equivalent |
|---|---|---|
| Standard | $25 per user per month | $20 per user per month |
| Premium | $125 per user per month | $100 per user per month |
OpenAI’s Business billing guide lists those rates. A workspace requires at least two paid seats, and Standard and Premium can be mixed. The annual equivalent is billed annually; it is not a month-to-month $100 Premium offer.
For example, one Standard seat plus one Premium seat costs $150 a month on monthly billing. At the annual rates, the same combination totals $1,440 for a year: ($20 + $100) × 12. Two Premium seats total $250 a month on monthly billing, or $2,400 a year at the annual rates. These examples exclude additional usage purchases and applicable taxes.
Only workspace owners can purchase or change paid seat types. OpenAI’s Business FAQ confirms that Business and API billing remain separate. The optional promotional API grant described below is a separate benefit requiring eligibility and activation.
For a small company, mixed seats can make more sense than giving everyone the most expensive option. A person who runs long agent tasks may need Premium capacity; a teammate who mainly reviews results may have a different requirement. Compare the workspace’s actual needs before making every seat the same.
The Business models and limits guide treats individual Chat model allowances separately from Work and Codex capacity. Do not assume that every model, image tool or other feature has received the same increase, or turn an older seat multiplier into a new fixed message count.
October 28 API credits are a separate benefit
The allowance increase reported here is an October 29 change, independently confirmed by the official release notes and Pro documentation. The October 28 API-credit benefit already discussed in our Terra article is different: eligible Pro and Business subscribers must activate it for a linked API organization. It does not increase ChatGPT or signed-in Codex usage limits and does not cover Codex usage through API keys. See the official eligibility, activation and credit terms. Do not add an API grant to a signed-in Work/Codex allowance when comparing capacity.
A small demo you can record without upgrading
Use your current subscription and a model already available to your account. Start by recording the allowance displayed in Settings → Usage, including whether the meter shows used or remaining capacity and when it resets. Keep account identifiers and billing details out of the recording.
Choose one model, one reasoning level and Standard speed, and keep them fixed. Run five separate requests with the same supplied data, using the same client and fresh conversations. Pause other Work/Codex activity during the measurement so background consumption does not distort the comparison. Record those settings beside the results.
Shared data: Product A: 3 units at $20; Product B: 2 units at $35; Product C: 5 units at $8. For each request, supply these same rows and request an answer based only on them.
Five objective checks: (1) calculate total revenue—$170; (2) calculate total units—10; (3) identify the product with the highest revenue—B, $70; (4) rank products by revenue, descending—B, A, C; (5) return a CSV with product and revenue columns—A/60, B/70 and C/40. Decide the required answer format before running and count a request as successful only when both its result and format are correct.
Record allowance before and after, elapsed time, successful completions out of five and any retries. If the dashboard is too coarsely rounded to show a difference, report consumption as below the displayed resolution. Do not turn an unchanged percentage into a claim of zero usage.
This is a repeatable recording protocol, not a completed Kingy benchmark. It shows how a small workload behaves under your current allowance. Demonstrating a measured twofold improvement would require a comparable earlier baseline; today’s entitlement change alone does not prove twice as many successful completions.
Work, Codex and ordinary Chat have different limits
Work and Codex share an allowance. Pro 100, Pro 200 and Pro 500 currently have no five-hour usage cap in those experiences, but included usage still runs out. Weekly limits and model-specific rules remain relevant. Check Settings → Usage for the remaining balance and reset times. OpenAI explains the usage windows here.
The separate weekly GPT-6 Pro limit in ordinary Chat is not universally doubled: Pro 100 and Pro 500 keep their current limits; lower-limit Pro 200 accounts move up to match higher-limit accounts. Source: Pro guide, Chat message limits.
GPT-6 Pro in Chat is powered by Astra, but access to Astra in Work does not give every plan the same Chat entitlement. Plus includes Astra in Work and Codex, while ordinary Chat’s Pro option has separate eligibility. A Pro subscription also does not mean unlimited use of the GPT-6 Pro model. See OpenAI’s model availability and Chat limits guide.
A useful way to audit your account is to record the product, model, effort setting and limit message together. “I ran out of ChatGPT” is less useful than identifying whether a particular Chat model stopped or the shared Work and Codex pool was exhausted. That distinction determines whether another available activity can continue.
Why double the allowance may not mean double the finished work
Allowance consumption varies with model choice, context, reasoning, tool use and the size of the task. OpenAI’s usage guidance explains that apparently similar requests can consume different amounts.
Consider two illustrative jobs. One asks for a short rewrite of a supplied paragraph. The other asks an agent to research a market, inspect several files, build a spreadsheet and correct its own mistakes. Counting both as “one prompt” hides most of the work.
If you keep the same model, settings and type of task, more allowance provides additional room. If you use the extra capacity to run longer jobs or more demanding settings, you may consume it quickly. That is a change in how you use the subscription, not proof that the increase failed.
Model efficiency is another variable. Readers evaluating routine computer work can also see our GPT-6.1 Terra guide. For your own comparison, repeat a representative task and measure the quality of the result, how much correction it needs and the allowance consumed. A cheaper incomplete answer may cost more time overall.
Pro 500 and Ultrafast: speed has a separate usage cost
Ultrafast remains a distinguishing feature of Pro 500. OpenAI’s pricing documentation specifies that GPT-6 Astra and GPT-6.1 Sol Ultrafast consume included subscription usage at 8× the Standard rate. When using purchased credits, the rate is 6× Standard. Fast mode uses included allowance at 2.5× Standard and purchased credits at 2×, where supported.
Those are usage-consumption multipliers, not promises that every task finishes that many times faster. A larger included pool can still disappear quickly when you consistently choose the most expensive speed mode.
For example, think of an imaginary pool containing 800 units. At a one-unit Standard cost, it supports 800 equivalent units of activity. At an eight-unit cost, it supports 100. This is an illustration of the multiplier, not a published OpenAI allowance or task estimate.
The business question is whether faster interaction saves enough time to cover the extra monthly spend and capacity consumed. Interactive debugging may benefit more from lower response latency than a research job you are happy to leave running while you work elsewhere.
What about dots and their extended limits?
Sottiaux’s announcement also says OpenAI will continue the extended limits for dots for now. That wording does not establish a permanent entitlement or a new end date.
OpenAI’s dots documentation explains the important boundary: conversations with a dot do not count toward ChatGPT usage limits, while tasks the dot starts or manages in Work or Codex count toward those products’ limits. Deeper work has its own plan allowance. Access also depends on rollout, plan and region.
For an ongoing assistant, separate metering does not mean every task it can trigger is free of other limits. If your dot delegates a substantial assignment, check the destination product’s usage as well as the dot’s own status. That will give a clearer picture of where capacity is being spent.
Can you buy additional usage instead of upgrading?
Eligible personal accounts can purchase credits for supported features after their included allowance is used. OpenAI’s personal-plan credit guide says these can cover eligible Codex, Work and Office-related features, depending on the account. They are separate from API credits and from the ChatGPT wallet. Purchased credits generally expire twelve months after purchase.
That creates a useful comparison for irregular users. A month with one unusually large project may call for a different decision from a job that exhausts the allowance every week.
Record the last month’s additional usage spending and compare it with the price difference to the next plan. Also account for speed access and other features you actually need. Credits can extend supported usage; they should not be treated as a universal way to unlock every higher-tier feature.
How to check the change on your account
First, confirm the subscription and workspace. Open your plan settings and check whether you are using a personal Pro account or a Business workspace. Similar names do not necessarily mean the same billing arrangement.
Second, inspect Settings → Usage. Read the label on each meter, the remaining allowance and its reset date. Compare the same product and model rather than a Chat message allowance with a Work budget.
Third, try a familiar task. Use the model, speed and effort settings you normally choose. A representative piece of work is a more useful comparison than an unusually large stress test.
Fourth, reassess any planned upgrade. If you were paying more solely to avoid the Pro 200 cut, the premise of that decision has changed. If your current plan still interrupts essential work, evaluate the next tier against the actual gap.
An active Pro subscription does not need to be canceled and repurchased to receive the updated allowance. OpenAI says the update is automatic.
Kingy’s take: reassess the capacity you now have
The useful outcome is a clearer buying decision. Existing high-allowance Pro 200 users keep the capacity they expected to lose. Other affected subscribers receive more room for Work and Codex. The right next step is to evaluate that updated capacity before choosing a more expensive subscription.
Our recommendation is to make the next billing decision with a short usage log: completed work, interruptions, extra credits and time spent correcting results. A plan earns its place when it helps you finish worthwhile work consistently. The revised allowances deserve to be judged on that result.
Sources and reporting notes
This article is based on OpenAI’s current Pro terms, Work and Codex pricing, usage guidance, Chat model documentation, Business billing, Business FAQ, Business model limits, credit terms and dots guide, checked October 29, 2026.
The Business Premium and extended-dot-limit announcement is attributed to Sottiaux’s October 29 X post; a publicly indexed copy was available during reporting while direct X retrieval was blocked. The Pro changes are independently confirmed in OpenAI’s updated Help Center. Historical Plus multipliers are presented only as context from the earlier reporting, not as newly verified quotas.
Illustrative photograph: Rolf van Root / Unsplash. The photograph shows an earlier ChatGPT interface, not the updated subscription dashboard.
Trending on Kingy
Keep reading with the stories getting the most attention now.
The Kingy Brief
Get The Kingy Brief.
AI changes, original tests and one practical thing to try. Fridays at 09:00 Vancouver time.
Free · Double opt-in · Unsubscribe anytime
Sign up for The Kingy Brief on Beehiiv.
Signup help and newsletter schedule
Signup form provided by Beehiiv. After submitting, check your inbox for "Confirm your subscription to The Kingy Brief" and open its confirmation link. Check Spam or Promotions if you cannot find it.
Fridays at 09:00 Vancouver time: source-checked AI changes, original tests and one practical thing to try. We skip a week when there is not enough verified material. Free. Unsubscribe anytime.