OpenAI doubles ChatGPT Pro limits and scraps the Pro 200 cut

What this is about

On October 29, 2026 (local time), OpenAI made a U-turn on usage limits for its paid ChatGPT plans. Thibault Sottiaux, who is responsible for ChatGPT and Codex at OpenAI and known online as "Tibo", wrote on X that the company is doubling usage limits on all Pro and Business Premium plans. At the same time, OpenAI is withdrawing the cut to the $200 "Pro 200" plan that was due to take effect the next day, October 30. ITmedia and Cryptobriefing report this.

How the cut came about

The backstory goes back to late summer. According to Android Authority, OpenAI reintroduced five-hour limits for Plus subscribers on ChatGPT Work and Codex on August 25. These are time windows in which only part of the weekly allowance can be used before you have to wait. Sottiaux justified this at the time with better resource management, and with the fact that casual and new users often used up their weekly limit by accident. Pro plans were initially exempt, though five-hour limits could come to them in a few months.

According to ITmedia, on September 10 OpenAI paused new sign-ups for Pro 200 because of high demand for "GPT-6 Astra", and reopened them on September 29. The calculation changed at the same time: measured in API prices, the limit would be halved, Tibo wrote, but the model had become more efficient and could get more work done than a month earlier. Existing customers received an email. There was criticism, but also voices who felt the efficiency gain.

What was announced at DevDay

The Next Web reports on DevDay 2026 in San Francisco, with 2,500 developers and more than 20 announcements. On pricing:

• Pro 500: a new plan at $500 a month, the only one with "Ultrafast". According to OpenAI, this mode generates up to 300 tokens per second in Codex, up to eight times faster than normal. (A token is, roughly speaking, a chunk of text a model works with.)
• Pro 200: From October 30, the allowance in ChatGPT Work and Codex was to drop from 20x to 10x the Plus allowance. Weekly GPT-6 Pro messages in chat were to fall from 200 to 100.
• Transition: Existing customers kept their limits until October 29 and were to receive a one-time credit worth $2,500, expiring at year's end. OpenAI did not want to bring back a five-hour limit for Pro 200.

Other announcements included shared team workspaces ("ChatGPT Space"), a meeting-notes tool for the Mac, @ChatGPT in Slack and Microsoft Teams, cloud tasks in Codex, and the continuously running agents called "dots". Dots are initially unavailable to Pro users in the EEA, Switzerland and the UK.

The reversal in detail

According to ITmedia, the move is part of a campaign Tibo started on October 5: for 28 days, there will be one clear product improvement every day, or the limits get reset. The reversal and doubling was the improvement for day 25. He described user criticism as "warm feedback". In Japan, the plans cost ¥16,800 (Pro 100), ¥30,000 (Pro 200) and ¥84,000 (Pro 500) per month, according to ITmedia.

According to Cryptobriefing, the dots agents get their own expanded limits, separate from the plan allowances. That lets OpenAI price and manage this usage separately.

Our take

For paying users, this is clear relief, especially heavy Codex users. The episode shows how sensitive subscribers are to hidden cuts: the halving was sold as "more efficient, so equivalent", which many weren't convinced by. It also remains unclear how OpenAI will pay for the higher limits. The sources say nothing on that. Whether the originally promised $2,500 credits still apply isn't documented either. Keep in mind that some figures come from vendor statements and press reports, and OpenAI can change the limits again at any time. The five-hour limits for Plus remain in place.