OpenAI Raises GPT-5 Thinking Limit to 3,000 Messages

OpenAI has taken a significant step by dramatically increasing GPT-5 Thinking usage limits for ChatGPT Plus users. The change was announced directly by OpenAI CEO Sam Altman via X on August 11, 2025. Previously, Plus users could only access 200 messages per week for this reasoning model, but now that figure has surged to 3,000 messages per week—a 15-fold increase.

This move aims to address the surge in demand for GPT-5 Thinking, which is designed to handle complex problems with deeper reasoning processes. Altman emphasized that, in addition to the expanded reasoning limits, all model classes will soon have higher limits than they did before GPT-5 was launched.

Increased Limits for All Models

The upgrade isn’t limited to GPT-5 Thinking. For GPT-5 standard, Plus users now enjoy 80 messages every 3 hours, equal to 640 messages per day or about 4,480 messages per week. This is well above the average daily usage of most users, offering greater flexibility for those who rely on ChatGPT for intensive work.

Altman also revealed that OpenAI will soon roll out a user interface (UI) update to display which model is currently active. This transparency feature will help users monitor their usage and plan accordingly.

Surge in Reasoning Model Adoption

OpenAI’s internal data shows a sharp rise in the percentage of users leveraging reasoning models. For free-tier users, usage grew from less than 1% to 7%. For Plus subscribers, the jump was even more notable, from 7% to 24%. These figures highlight that more people are turning to GPT-5 Thinking for deep analysis and multi-layered logical work.

With the weekly cap raised to 3,000 messages, Plus users now get about 428 GPT-5 Thinking messages per day. While still below the GPT-5 standard’s weekly capacity of 4,480 messages, the change represents a major step forward in unlocking the model’s full potential for regular users.

Impact on Professional Users

For Pro and Team subscribers, OpenAI continues to offer virtually unlimited access to GPT-5, including the Thinking variant. The only restrictions are usage policies to prevent abuse. This gives a significant advantage to enterprises, researchers, and developers processing large volumes of data.

For Plus users, managing message allocation becomes more strategic. With a more generous weekly limit, they can prioritize high-precision tasks such as data analysis, strategic planning, or creative content development.

Focus on Capacity and Quality

Altman noted that in the coming days, OpenAI will share plans for managing capacity over the next few months. This will involve balancing ChatGPT versus API usage, prioritizing existing versus new users, and allocating between research and product development. This approach is expected to maintain service quality as the user base grows.

With this new policy, OpenAI demonstrates its commitment to meeting user needs without compromising performance. The increase of GPT-5 Thinking to 3,000 messages per week is not only good news for active users but also a signal that reasoning models will be a key pillar in AI’s future ecosystem.

This strategic move shows that AI innovation doesn’t just come from improving model intelligence—it also comes from implementing more flexible and transparent usage policies. For those looking to maximize GPT-5 Thinking’s capabilities, now is the time to integrate AI on a broader and more effective scale.