OpenAI doubles ChatGPT Pro usage limits for all tiers

Reversing the usage cut

OpenAI has officially announced that it will not be reducing the usage limits for its high-end ChatGPT Pro plans. This decision comes as a direct response to feedback from customers who expressed dissatisfaction with the company's original proposal. The company had previously intended to lower the included usage allowance for existing subscribers on the $200 per month plan starting on October 29.

Instead of moving forward with the planned 50% reduction, the company decided to take the opposite approach. They are now doubling the usage limits across all Pro tiers, which include the $100, $200, and $500 price points. This reversal ensures that high-paying users maintain their ability to perform large volumes of work without facing unexpected restrictions.

The announcement was shared via social media by Tibo, who serves as a primary representative for ChatGPT and Codex. He confirmed that these changes are effective immediately. By doubling the limits rather than cutting them, OpenAI is attempting to maintain goodwill with its most dedicated and highest-spending user base.

Impact on Work and Codex

The updated usage allowances primarily affect the ChatGPT Work and Codex tools. These two services share their usage allowances under eligible subscriptions, meaning the increase in limits provides more breathing room for professional workflows. Users who rely on these specific tools for coding or complex business tasks will see a significant increase in their available capacity.

Because the limits are shared, the doubling of the allowance allows for more flexibility in how users allocate their resources. Subscribers can now perform more actions across both platforms without worrying about hitting a ceiling as quickly as they would have under the previous plan. This is particularly important for teams that integrate these tools into their daily production cycles.

While OpenAI did not provide the exact numerical limits for each specific tier, the move is clearly designed to provide more value. The goal is to allow Pro subscribers to utilize Work and Codex more extensively without being forced to upgrade to a more expensive subscription. This keeps the current pricing structure viable for heavy users who were on the fence about the planned cuts.

Dots and subscription independence

In addition to the changes for the Pro tiers, OpenAI clarified the status of Dots, which are the company's always-on AI assistants. The company confirmed that they are continuing to offer extended usage limits for Dots for the time being. This ensures that users who rely on these specific assistants do not see a change in their experience alongside the broader Pro tier updates.

A key technical detail of this update is that the usage of Dots will remain separate from the default usage of the subscription. This independence allows users to interact with their assistants without depleting the primary limits used for ChatGPT Work or Codex. It creates a distinct lane for assistant-based interactions, preserving the core limits for other types of tasks.

By keeping these limits separate, OpenAI provides a more modular experience for its power users. Subscribers can manage their primary work tasks and their assistant interactions as two different types of activity. This distinction helps in planning how much of the subscription's

Security insights from the Blue Report

The Blue Report 2026 provides a detailed look at how defenses perform against various techniques. It measures these defenses across 338 million simulations conducted in customer production environments. This data helps identify how systems hold up under real-world pressure rather than just in controlled tests.

One significant finding involves how security measures behave once a breach occurs. When attackers obtain valid credentials, only 37% of their actions are blocked by existing defenses. This indicates that overall prevention scores can sometimes hide the reality of what happens after an initial access point is compromised.

The report highlights that protection levels drop sharply once an intruder is using legitimate credentials. Because the system recognizes the user as authorized, many internal barriers are bypassed. This underscores the importance of monitoring behavior after the initial login rather than relying solely on perimeter security.