OpenAI Doubles Pro Codex Limits and Scraps the Pro 200 Cut

OpenAI doubled ChatGPT Work and Codex usage allowances on its Pro and Business Premium plans and cancelled a planned cut to Pro 200, effective immediately.

The change was announced on 29 October 2026 by Thibault Sottiaux, who leads Codex and ChatGPT Work, in a post on X, after customer feedback on the planned reduction. The write-up that compiled the details is explicit that it relays reporting from BleepingComputer, ghacks and xenospectrum rather than first-hand testing, and that OpenAI does not publish exact numerical usage limits per tier.

What changed

• Pro 100 and Pro 500 allowances double — Work and Codex usage on the $100-per-month and $500-per-month plans doubles at the same prices.
• The Pro 200 cut is dead. OpenAI had planned to drop the included Work and Codex allowance for $200-per-month subscribers from 20x the Plus allowance to 10x Plus, and to cut GPT-6 Pro chat messages from 200 to 100 per week. That is off the table, and all Pro 200 subscriptions are levelled to the same allowance regardless of signup date.
• Business Premium gets the same doubling, also at unchanged prices.

No subscription price moves in either direction. This is a limits change, not a pricing change.

What did not change

The weekly GPT-6 Pro message cap inside regular Chat stays where it was — the doubling applies to Work and Codex only. ChatGPT Plus limits are untouched; Plus subscribers get nothing from this announcement. Dots, OpenAI's always-on assistants, keep their own extended limits, and Dots usage stays separate from the Work and Codex allowance, so using one does not draw down the doubled pool. The reporting notes those extended Dots limits are described as continuing "for now", with no end date attached.

The API credits are a different pot

A day earlier, OpenAI launched API credits for eligible subscriptions. Simon Willison, who holds subscriptions with both vendors, describes it as near-identical to a benefit Anthropic announced earlier this month for Max and Team: subscribers on $100, $200 or $500 plans receive the equivalent quantity of credits to apply to an API organization — expiring credits, in the Anthropic version he describes.

Those credits apply to API spending. This week's announcement applies to Work and Codex usage inside the subscription. Neither extends the other, and conflating them is the easiest way to over-estimate what a Pro seat now buys.

Is Pro worth it now?

If you were rationing Codex sessions near the end of a billing cycle on Pro 100 or Pro 500, this is the announcement that matters: twice the room, same invoice. If you were on Pro 200 and planning to downgrade in protest, the thing you were reacting to no longer exists.

If you are on Plus, nothing here moves. If you never approach a cap, you will not feel this at all, and the extra allowance is not a reason to upgrade — unmeasured headroom you do not use is worth nothing. And if your bottleneck is GPT-6 Pro messages in regular Chat rather than Codex throughput, this announcement does not address it.

What to watch

Because OpenAI publishes no per-tier numbers, the doubling cannot be checked against any public table — you will only see it in your own usage over the coming weeks, and any figure quoted as a precise task or message count is invented. Worth logging your own cap-hit frequency before and after, since that is the only measurement available to you.

Also worth noting the pattern: Pro limits have now moved twice in quick succession, first a planned cut, then a doubling under pressure. Treat the current allowance as this month's allowance, not a commitment, and avoid architecting a workflow that only pencils out at the doubled rate.