A month ago, OpenAI told a slice of its most expensive customers that their usage was about to shrink. On October 29, 2026, it un-shrank it.

The short version: the Pro 200 allowance reduction announced in OpenAI's September 29 email is canceled, and instead of anyone losing usage, every Pro tier's Work and Codex allowance went up. If you pay for Pro, you got more yesterday and your price did not move. Here is the longer version, because the details differ by tier and the fine print has a few traps.

How we got here

On September 29, OpenAI introduced Pro 500, a new $500-a-month tier whose headline perk is Ultrafast mode for GPT-6 Astra and GPT-6.1 Sol. The same announcement said Pro 200's included usage allowances were changing, with affected users to be notified by email. Read by most people as: the $200 plan was about to get less generous while a shinier $500 plan took the top spot.

The reaction was loud enough that OpenAI's Codex lead Thibault Sottiaux acknowledged it directly when announcing the reversal: the company heard the feedback, and Pro 200's limits are not being reduced after all.

What changed on October 29, tier by tier

Pro 100, the $100-a-month plan: the Work and Codex allowance doubled from its previous level.

Pro 200, the $200-a-month plan: this is where the nuance lives. Pro 200 subscribers had been split into two allowance levels. If you already had the higher Work and Codex allowance, you simply keep it—no reduction, no doubling. If you were on the lower allowance, yours doubled starting October 29, which brings you up to the same level as everyone else on the tier. Net effect: one Pro 200 allowance, equal to the higher of the two old ones.

Pro 500, the $500-a-month plan: doubled Work and Codex allowance, same as Pro 100.

Business Premium plans got the same doubling treatment, per OpenAI's release notes.

What did not change

Several things, and they matter for anyone about to celebrate too early.

Plus is untouched. The $20 plan's usage stays exactly where it was.

Prices did not move anywhere. Pro still runs $100, $200, and $500 a month, billed monthly, with no annual option.

The weekly GPT-6 Pro message limits in Chat are a separate system from the Work and Codex allowance. Pro 100 and Pro 500 keep their current weekly chat limits; on Pro 200, the lower weekly limit now matches the higher one. So this is not "every chat count on every plan just doubled"—the doubling is specifically about Work and Codex usage, with a chat-limit alignment on Pro 200 only.

Ultrafast is still exclusive to Pro 500. The expanded Pro 200 allowance does not unlock it, and buying credits on Pro 100 or Pro 200 does not either. For the record, Ultrafast draws from the included allowance at eight times the standard rate, then from purchased credits at six times the standard credit rate—speed costs extra by design.

And one honest caveat from OpenAI's own help pages: a doubled allowance does not guarantee twice as many completed tasks. What a task consumes still depends on the model, context length, reasoning effort, tool calls, and speed mode. Fast mode, for instance, eats the allowance at 2.5 times the standard rate.

Do you need to do anything?

No. Existing subscribers receive the updated allowance automatically—OpenAI is explicit that you do not need to wait for renewal, and you definitely should not cancel and resubscribe in hopes of triggering something. If you canceled in protest earlier and your subscription is still in its paid period, it remains active until the period ends, and the cancellation can be undone before then.

If you let your Pro 200 lapse entirely during the drama and want back in, that door is open too: Pro 200 is available for new subscriptions again, at the same $200, including the updated allowance.

Why it is worth a second look

Allowance math is invisible until it hits you, which is why plan changes like this generate so much heat. The interesting read on this reversal is not just that OpenAI flinched. It is that the Pro tiers have quietly become a three-rung ladder—$100, $200, $500—with the differentiation shifting from raw usage toward speed tiers and exclusives like Ultrafast. Usage allowances, the thing subscribers actually watch, just got more generous across the board rather than narrower at the top. For anyone who was mid-argument with themselves about downgrading, the decision just got easier: the plan you have is now worth more than it was on Tuesday.