OpenAI CEO Sam Altman has announced a temporary increase in rate limits for all reasoning models, benefiting both free and paid users. The change, which applies across every model, comes in response to growing demand but may lead to some feature restrictions for free-tier users. In a post on X (formerly Twitter), Altman confirmed, “All model-class limits will shortly be higher than they were before GPT-5.”
Why the Change?
The move follows criticism over GPT-5’s reasoning performance and tone. Over the past weekend, many users complained the AI’s conversational style felt shorter and less warm compared to earlier versions. This feedback prompted OpenAI to restore GPT-4 access for some users, though GPT-5 now remains the unified standard model, combining both GPT- and o-series models.
ALSO READ: Microsoft Just Made 3D Creation Faster And Easier: Copilot 3D Launched
Rising Usage Numbers
Altman revealed a sharp jump in usage:
- Free-tier users: Daily usage rose from 1% to 7%.
- Plus subscribers: Usage surged from 7% to 24%.
Free users currently have access to the GPT-5 Thinking model, while paid subscribers can choose between multiple reasoning models, including GPT-4o.
ALSO READ: Meta Hires From Apple Again: Pulls Out More AI Workers As Talent Hunt Continues
Balancing Performance and Costs
Boosting rate limits for computationally heavy reasoning models is costly. Altman admitted the company will need “capacity trade-offs” to sustain the change. OpenAI will evaluate its options in the coming days to balance cost, performance, and accessibility.
The Future of Reasoning in AI
Altman expects reasoning capabilities to become a central part of AI’s future. “I expect use of reasoning to increase over time greatly, so rate limit increases are important,” he noted. For now, users on all tiers can enjoy expanded access—though it may be temporary.