On October 29, 2026, OpenAI canceled the planned October 30 reduction of limits for the $200/month ChatGPT Pro 200 tier: the high limit for ChatGPT Work and Codex is preserved, subscribers with reduced limits have their volume doubled to the same level, and the Pro 200 subscription is open for new sign-ups again. Included ChatGPT Work and Codex usage is also doubled on the $100/month Pro 100 and $500/month Pro 500 tiers, while prices for all three plans remain unchanged. Increased limits are already active for existing subscribers.

What happened

The decision was reversed one day before it was set to take effect. On the Pro 200 tier ($200/month), the high weekly limit for ChatGPT Work and Codex is preserved, subscribers with reduced limits have their volume doubled to the same level, and the subscription is open again for new customers. On Pro 100 ($100/month) and Pro 500 ($500/month), included ChatGPT Work and Codex usage is doubled, and prices for all three plans remain the same. Increased limits were applied automatically, without waiting for a new billing cycle or re-subscribing. Weekly message limits for GPT-6 Pro in standard chat did not change on Pro 100 and Pro 500; on Pro 200, the reduced weekly limit was aligned to the high level, and Plus limits remained unchanged. The Pro 500 tier received Ultrafast for GPT-6 Astra and GPT-6.1 Sol: this mode consumes the included limit at 8x the speed of Standard and is not available on Pro 100 and Pro 200, even with purchased credits. Changes are confirmed by the official OpenAI Help Center article "About ChatGPT Pro tiers," updated on October 29, 2026.

Context

The event has a backstory: a letter dated September 29 announced a limit reduction on Pro 200 that was set to take effect on October 30, meaning OpenAI made two opposite decisions regarding the same quotas in less than a month—first to cut, then to roll back and expand. ChatGPT Pro tiers are divided into three price levels, and ChatGPT Work and Codex are the agentic work components of the subscription around which the changes were built. Both official OpenAI Help Center articles—"About ChatGPT Pro tiers" and "ChatGPT Work and Codex"—record the new limits, but the motives for the reversal are not disclosed: versions suggesting that calculated capacity or inference economics allow keeping more included work remain unconfirmed hypotheses. There is no scientific novelty in the event—no new model, benchmark, or methodology; the changes concern only pricing policy.

Why this matters for the industry

For the industry, this is a shift in priorities: instead of planned margin savings, OpenAI is retaining its most expensive subscribers by giving them twice as much included agentic work at the same prices. For product builders, this expands the "agent budget" within a single subscription: if the quotas hold, agentic pipelines on Codex will become more viable to sell as a service with a fixed subscription price rather than variable API costs, and a wave of workflow templates may emerge around Codex. The Ultrafast speed mode, locked exclusively to the Pro 500 tier, shows how the top price layer can monetize speed and priority unavailable to lower tiers, although this is an interpretation rather than a stated strategy. The main lesson for companies from this reversal: quotas are a variable, not a commitment, so product plans should maintain a fallback path via API in case of another limit review.

Why this matters for users

There is nothing to do: increased ChatGPT Work and Codex limits are already working, and there is no need to re-subscribe or wait for the next billing cycle. Usage and remaining limits can be checked in Settings → Usage, and in Codex—under Usage & Billing. The one practical step now is to review which batch tasks Codex now fit within your tier: with the doubled limit, long agentic runs will not hit the quota as quickly. Pro 200 subscribers whose limits were previously reduced now have access to the restored top service level: double the work volume and a high weekly limit for GPT-6 Pro messages in standard chat. Pro 100 and Pro 200 subscribers should not expect Ultrafast: purchasing credits will not unlock this mode; it remains a Pro 500 privilege, and weekly chat limits on Pro 100 and Pro 500 remained unchanged.

What is still unknown / limitations

The main unknown is the unit of work: sources state that nominal ChatGPT Work and Codex limits are doubled, but it is not defined what counts as one unit, so the claim that the effective price per unit of agentic work has dropped by approximately half is not yet confirmed. The motives for the reversal are not disclosed: versions about freed-up capacity or changed inference economics remain hypotheses, as is the interpretation that OpenAI is betting on long agentic scenarios rather than standard chats. The idea that the doubling signals a decrease in the unit cost of GPT-6-class inference is also speculative: OpenAI has not disclosed any timelines or plans. Finally, the history with the September 29 letter shows that quotas can be reviewed in both directions—the company has not guaranteed that the current doubled limits will remain in place.

Sources

Author

Look at AI, editorial team