OpenAI Announces Increased Rate Limits And UI Enhancements For ChatGPT Plus Following GPT-5 Release

In Brief

OpenAI is addressing GPT-5 rollout issues by improving model auto-switching, adding a manual thinking trigger UI, doubling Plus user rate limits, and considering the return of GPT-4o.

CEO of the artificial intelligence research organization OpenAI announced that the team is increasing rate limits for reasoning capabilities available to ChatGPT Plus users, with all model-class limits expected to soon exceed those in place before the release of GPT-5. Additionally, a user interface update will be implemented shortly to clearly indicate which model is currently in use. It is also planned to allow for users to manually initiate advanced reasoning functions.

Additionally, OpenAI noted that the percentage of users utilizing reasoning models on a daily basis has grown substantially; for example, the proportion of free users engaging with these models rose from less than 1% to 7%, while among Plus users it increased from 7% to 24%. The expectation is that the use of reasoning features will continue to grow over time, making the increase in rate limits a necessary adjustment.

Following the release of ChatGPT-5 is OpenAI’s latest AI model last week, the team acknowledged some unexpected difficulties during the process, describing the experience as “bumpier” than anticipated due to multiple simultaneous updates. Specific issues mentioned included a period when the autoswitcher was not functioning, which negatively affected GPT-5’s performance, making it seem less capable than usual. According to the team, efforts are being made to improve the system to ensure users access the appropriate model more consistently, alongside increasing transparency about which model is responding to each query.

The rollout to all users is taking longer than initially planned, as it represents a large-scale change. OpenAI is also exploring options to allow Plus users continued access to GPT-4o while gathering data on the potential trade-offs involved.

ChatGPT-5: Advanced Reasoning, Adaptive Model Selection, Enhanced Performance, And Seamless Integration For Superior AI Interaction

ChatGPT-5 introduces advancements in reasoning, speed, and task adaptability. It features a unified system that intelligently selects the optimal model configuration—standard, mini, or nano—based on task complexity, eliminating the need for manual model selection. Users on free plans have usage limits, while Plus and Pro subscribers enjoy higher or unlimited access, respectively.

The model excels in various domains, including coding, writing, health, and visual perception. It offers improved performance in coding tasks, scoring 74.9% on SWE-bench Verified and 88% on Aider polyglot, surpassing its predecessor, GPT-4o. GPT-5 also introduces four new user-selectable personalities—cynic, robot, listener, and nerd—to enhance conversational dynamics.

Additionally, it integrates with external services like Gmail and Google Calendar, enabling users to manage scheduling and email tasks. It also reduces AI hallucinations by 45% compared to its predecessor, making it more reliable for users seeking factual and accurate responses.