OpenAI doubled the work limits on its ChatGPT Pro plans on October 29, 2026. The increase covers the $100, $200 and $500 per month Pro tiers and the business-focused Business Premium plan, and subscription prices stay the same. A planned reduction in allowance for the $200 plan was withdrawn at the same time.
| Item | Detail |
|---|---|
| What | Default ChatGPT Work and Codex usage limits doubled; planned cut to Pro 200 limits withdrawn |
| Who | OpenAI (announced on X by Tibo Sottiaux, who leads ChatGPT and Codex) |
| When | October 29, 2026 (local time), effective immediately |
| Price | Pro at $100, $200 and $500 per month and Business Premium all keep their current prices |
| What it means for you | Top-tier paid users get more agent work with no sign-up step; the Plus plan is unchanged |
What happened
Sottiaux, who oversees ChatGPT and Codex at OpenAI, announced the increase in a post on X. Most of the extra allowance applies to two products: ChatGPT Work, an agent for automating office tasks, and Codex, OpenAI’s coding agent. Subscribers on the three Pro tiers and on Business Premium do not need to apply or pay again. The larger limits are available right away.
The $200 Pro plan, known as Pro 200, had a slightly different history. In September OpenAI reopened new sign-ups for Pro 200 and said it would lower the amount of usage the plan provided. Existing subscribers were guaranteed the old limits until October 29. OpenAI has now backed away from the reduction in response to user feedback. People who signed up under the lower limits were brought up through the doubling to the same level as existing subscribers.
Not every kind of usage went up. According to OpenAI’s support documentation, the number of weekly GPT-6 Pro model uses in the regular chat window is unchanged for both Pro 100 and Pro 500. The lower-priced Plus plan also sees no change. The extra usage bonus that came with Dots, OpenAI’s personal-assistant agent, continues for now, and Dots usage is counted separately from the main ChatGPT limit.
Why it matters
AI agents increasingly work by taking one instruction and running for a long time. For people using them at work, how long an agent can keep running now matters as much as how capable the model is. Anyone who hands Codex a large coding job, or uses ChatGPT Work to automate email and calendar tasks, regularly hits a limit that stops the work partway through. If the default allowance doubles at the same subscription price, the effective cost of each job falls.
OpenAI also added a caveat: doubling the limit does not mean exactly twice as much work gets done. Consumption varies widely with the model you pick, the difficulty of the task and the speed option you choose. Pro users should check the real effect against their own work patterns.
Background
The increase is the 25th announcement in a 28-day run of daily updates that OpenAI announced on October 4. At the time, Sottiaux said that for four weeks OpenAI would ship an improvement Codex and Work users would notice each day, or give a reset that refills limits users had used up. During the campaign OpenAI raised the default speed of GPT-6 Astra and GPT-6.1 Sol by about 50%, and it also introduced GPT-6.1 Terra, a low-cost model aimed at automation.
New features arrived as well. A plugin connecting to iCloud Mail and Calendar was added, and Dots became usable over text messages. Early in the campaign the emphasis was on temporary resets for users who had hit a limit. The day-25 announcement is different because it enlarges the base allowance built into the plans themselves. After the announcement, Sottiaux said in effect that OpenAI had moved on from the stage of handing out resets to a stage of giving out tokens generously.
What to watch next
The 28-day campaign has not formally ended, so more announcements are possible over the remaining days. OpenAI appears set to keep pursuing two tracks at once: making models more efficient so the same limit does more work, and widening the base allowance. Sottiaux also signaled that better models and efficiency gains will keep increasing the amount of work users can actually get done.
For practical decisions, Pro 200 subscribers can carry on with their existing work plans without worrying about a cut. Pro 100 users may want to ask whether the bigger limits reduce their reason to move up to a higher plan. People on the Plus plan, or those who mostly use ChatGPT for ordinary conversation, are not direct beneficiaries of this change. Readers comparing top-tier plans across vendors may also want to read our piece on Anthropic’s Claude Ultra plan change.
