OpenAI CEO Sam Altman announced an increase in rate limits for reasoning models for all ChatGPT users, a temporary change affecting all model classes, driven by a surge in users accessing reasoning capabilities, with future capacity trade-offs yet to be detailed.
OpenAI CEO Sam Altman recently stated that the company is increasing the rate limit for reasoning models across all ChatGPT user tiers. This adjustment, while comprehensive in its application to all model classes, is explicitly temporary. OpenAI indicated that it will subsequently determine how to manage future capacity trade-offs, which suggests potential alterations for users on the free tier. This decision stems from new data indicating a rise in the number of users engaging with the reasoning capabilities offered by the AI models.
today we are significantly increasing rate limits for reasoning for chatgpt plus users, and all model-class limits will shortly be higher than they were before gpt-5.
we will also shortly make a UI change to indicate which model is working.
— Sam Altman (@sama) August 10, 2025
In a post on X, formerly Twitter, Altman confirmed that all ChatGPT users would experience higher rate limits for reasoning models. He further stated that “All model-class limits will shortly be higher than they were before GPT-5.” This phrasing is notable, as OpenAI had previously retired older models concurrently with the release of GPT-5. However, this policy was reversed over the weekend, when the company reinstated the GPT-4o artificial intelligence model following user feedback. Users had expressed dissatisfaction, noting that the conversational responses from the latest large language models were perceived as less engaging and shorter compared to the older models.
The current model structure means that free-tier users have access only to the GPT-5 Thinking model, which is integrated into the standard GPT-5 AI model, a unified system combining GPT- and o-series models. In contrast, paid subscribers retain access to a distinct Thinking model, which they can select via a model picker.
Additionally, GPT-4o, an o-series model, also possesses reasoning capabilities. Providing these elevated rate limits for reasoning models imposes significant operational costs on OpenAI, a company already operating at a loss due to substantial expenses. Altman acknowledged that managing these increased rate limits would necessitate “capacity trade-offs,” with further details on how the company plans to address this to be released in the coming days.
tomorrow or tuesday we expect to share our thinking on how we are going to make capacity tradeoffs over the coming months.
(e.g., chatgpt vs the api, existing users vs new ones, research vs product, etc)
— Sam Altman (@sama) August 10, 2025
Altman elaborated on the rationale behind the increased rate limits for reasoning models, attributing it to a demonstrable rise in the daily percentage of users accessing these specific capabilities.
the percentage of users using reasoning models each day is significantly increasing; for example, for free users we went from <1% to 7%, and for plus users from 7% to 24%.
i expect use of reasoning to greatly increase over time, so rate limit increases are important.
— Sam Altman (@sama) August 10, 2025
He provided detailed statistics: the percentage of free users utilizing reasoning models increased from one percent to seven percent. For Plus subscribers, the growth was even more pronounced, with usage rising from seven percent to 24 percent. Altman also commented, “I expect use of reasoning to increase over time greatly, so rate limit increases are important.”
