ChatGPT Pro Codex Limits Doubled: Who Actually Gets More?

Pro 100 and Pro 500 now include twice their previous ChatGPT Work and Codex allowance. Pro 200 users either keep their existing higher allowance or move from the lower allowance to that higher level. Check your current usage before upgrading: the Pro update applies automatically, and Plus usage is unchanged.

As of 2026-11-01, the key distinction is twice the previous allowance for some subscribers, rather than twice as much for every account. OpenAI’s October 29 release notes and current Pro tier guide explain the difference.

Which plans actually received more usage?

Your previous allowance determines whether this change adds capacity or preserves what you already had.

Plan or previous allowance	Change announced on October 29	What it means for you
Pro 100	Work and Codex allowance doubled	More included capacity at the same tier.
Pro 200, previously on the lower allowance	Lower allowance doubled to match the higher allowance	You move up to the allowance existing higher-allocation subscribers keep.
Pro 200, already on the higher allowance	Planned reduction canceled	You retain that allowance. It does not double again.
Pro 500	Work and Codex allowance doubled	More included capacity at the same tier.
Business Premium	Increase confirmed; Tibo’s announcement described doubling	The updated comparison with Standard seats is not established in the current Help Center pages.
Plus	Usage unchanged	This announcement adds no Plus allowance.

Two Pro 200 subscribers can now have the same allowance even though only one received an increase. The canceled reduction was the one announced in OpenAI’s September 29 email.

Business Premium needs more care. Tibo’s October 29 announcement says its limits doubled, while the release note confirms an increase without a multiplier. The Business FAQ still describes Premium as having 5x Standard’s usage. Do not turn that into a verified 10x entitlement by multiplying the two announcements.

Likewise, “twice the previous allowance” does not establish a current Pro-to-Plus ratio. Use it to understand your tier’s change, not to infer fixed message counts or a new context window.

When does the increase apply, and is it permanent?

The Pro increase is already effective; existing subscribers do not need to wait for renewal. OpenAI dates the adjustment to October 29, 2026.

The current Pro guide says existing accounts receive the updated allowance automatically. An allowance increase also does not establish that your consumed usage was reset to zero.

The October 29 announcement gives no expiration date for the Work and Codex increase. Treat it as the current included allowance. It is not a published guarantee that plan limits will never change.

The extended dots allowance has different wording: Tibo said it would continue “for now.” Keep that temporary commitment separate from the Pro allowance update.

What shares the Work and Codex allowance?

On the plans covered here, Work and Codex share an included pool. Research in Work can therefore leave less capacity for coding. OpenAI’s usage guidance also says local messages and cloud chats share your plan’s allowance.

A larger allowance does not promise twice as many completed tasks. Consumption depends on the model, task, and settings. Comparing a short edit with a long repository investigation will not tell you whether your entitlement changed.

As of 2026-11-01, Pro plans have no five-hour Work/Codex limit. Business Premium also has no five-hour limit, according to the Business limits guide. Included capacity and applicable weekly limits still matter.

For Plus or Business Standard, distinguish the five-hour window from the weekly allowance. Where both apply, you need capacity in both. A later weekly reset does not make an exhausted five-hour window usable.

Also include participating apps in your review if you authorized them to use your ChatGPT plan. Those requests can consume the same allowance. Our guide to ChatGPT plan usage in partner apps explains that separate authorization and billing boundary.

Ordinary Chat requires a date distinction. OpenAI’s broader ChatGPT/Codex sharing update is scheduled to start in December, with Instant chat excluded. As of this article’s date, do not treat that future change as already applied to every conversation.

Does dots usage count toward these limits?

Conversations with your dot do not consume your normal ChatGPT allowance, but delegated Work and Codex tasks do.

The dots documentation explicitly distinguishes those activities. If your dot starts or manages a coding task in Codex, that task follows Codex’s normal usage limits. The same applies to Work tasks.

Tibo’s extension of separate dots limits does not make those delegated tasks unlimited. When investigating unexpected consumption, check the tasks your dot started as well as tasks you started directly. The dots limits and permissions guide covers the operational distinction.

How should you check your account before paying more?

Check the active account, exhausted allowance, and reset time before changing plans.

1. Open Settings → My Plan and confirm your subscription.
2. Open Settings → Usage and record the remaining allowance and displayed reset times.
3. In an active Codex CLI session, enter /status to inspect its usage information.
4. Compare a few representative tasks with the same model, reasoning level, speed setting, and acceptance criteria.

OpenAI documents the account checks in Using Codex with your ChatGPT plan. For Business, also confirm which workspace and seat type you are using.

If the allowance still looks wrong, contact support with the affected limit, reset time, client, model, and time zone. Include a screenshot with sensitive information removed. A percentage alone cannot show whether the underlying allowance changed.

Review your model, reasoning level, and speed setting before buying capacity. Use OpenAI’s Work and Codex usage guide to choose settings for a bounded task whose result you can check. Compare the result and allowance used before increasing effort.

Which subscription should you choose now?

Keep your current tier if the updated allowance covers your normal work. Choose a higher tier for a recurring capacity or feature need.

The current individual prices are USD $100/month for Pro 100, $200/month for Pro 200, and $500/month for Pro 500. Those prices do not translate into a guaranteed number of tasks.

• Stay on Plus if its unchanged allowance is sufficient. An occasional deadline is a reason to compare overflow options before accepting a larger recurring bill.
• Start with Pro 100 if you need a Pro plan and want to evaluate its newly expanded allowance. Measure a normal working week before choosing a higher tier.
• Choose Pro 200 if Pro 100 repeatedly interrupts work that the higher tier can cover. If you already had Pro 200’s higher allowance, first assess the capacity you retained.
• Choose Pro 500 when its capacity or Ultrafast access justifies the recurring cost. Among Pro plans, Ultrafast remains exclusive to Pro 500; it is not added to Pro 200 by this update.

For teams, upgrade the teammates who need Premium rather than assuming every seat needs the same capacity. Business permits mixed Standard and Premium seats.

As of 2026-11-01, the Business list prices are $100 per Premium seat per month with annual billing, or $125 with monthly billing. Standard costs $20 or $25 respectively. The billing guide requires at least two paid seats.

For example, one Premium seat plus one Standard seat costs $1,440 annually, equivalent to $120/month. Monthly billing costs $150/month for that pair. This is USD list-price arithmetic before applicable taxes; checkout prices can vary by country and currency.

A solo developer should therefore compare the full workspace commitment with a personal Pro plan. An existing Business team should compare the cost of changing a seat within its current workspace.

When should you use credits or an API instead?

Use subscription credits when you need to continue within the signed-in experience. Consider an API when the overflow work fits local Codex or automation.

Continue within your subscription

On Plus or Pro, check for an option to add credits after reaching a Codex limit. If purchases are available for your account, ChatGPT usage credits let you continue using eligible features. Business teams can use purchased workspace credits when workspace settings and spending controls permit it.

These are separate from API billing. The subscription-linked promotional API benefit explicitly excludes Codex usage through API keys. Do not count that grant as funding for your Codex overflow budget.

Compare paid API overflow with the upgrade cost

API-key authentication charges Codex usage to the API account instead of the subscription allowance. OpenAI’s authentication guide documents API-key access for local work in the Codex CLI, ChatGPT desktop app, and IDE extension. Codex Cloud requires ChatGPT sign-in. API access still has its own rate and account limits. If you depend on hosted workflows, check the Codex client comparison before switching.

For occasional overflow, compare the work you would move to the API with the incremental subscription cost. Pro 100 to Pro 200 adds $100/month in base subscription charges. Pro 200 to Pro 500 adds $300/month.

Use the same model, reasoning effort, speed, task scope, and quality standard. Include every agent turn, repeated context, retry, and rejected attempt in the token totals.

For a concrete estimate, OpenAI’s API price table lists gpt-6.1-sol at USD $2.00 input and $10.00 output per 1M tokens. These are Standard-speed, short-context rates as of 2026-11-01.

Illustrative token subtotal: assume monthly overflow totals 8M uncached input tokens and 1M billed output tokens. Assume all requests qualify for short-context pricing, with no caching, paid tools, or regional-processing surcharge.

8 × $2.00 + 1 × $10.00 = $26.00

This is an estimate before taxes, not a measured Codex bill. Your actual bill must include any applicable cache writes, cached reads, tools, infrastructure, provider fees, and processing surcharges. Keep input billing categories distinct so cached tokens are not counted twice.

Compare that complete cost with the $100 upgrade increment while retaining Pro 100. Also check whether Pro 200 would cover the workload and whether both routes provide the features you need. The token subtotal alone does not establish an equivalent replacement.

A gateway offers another API route for local coding. BetterToken’s current Codex integration guide documents the official CLI with https://www.bettertoken.ai/v1 and wire_api = "responses". This is separate pay-as-you-go API access, not a bundled ChatGPT subscription. Check the selected model and current prices before budgeting.

For repeated work, compare the cost of an accepted result, including unsuccessful attempts. The cost-per-successful-Codex-task method explains how to structure that comparison. Record enough representative work to decide whether occasional API overflow or a higher recurring allowance fits your actual month.