OpenAI has completed the full rollout of GPT-5 for all Plus, Pro, Team, and Free users. The company restored access to GPT-4o following user complaints about the behavior of the new model. According to Sam Altman, OpenAI acknowledges that it did not account for the demand for the previous version and allows users to choose between models through settings.
Plus and Team users received doubled limits on GPT-5 usage until the end of the week. If someone reaches the message limit, the system automatically switches to the “mini” version of GPT-5 or to GPT-5 Thinking until the limits are renewed. Now these options are directly available in the main model selection menu for Pro accounts.
OpenAI continues to work on improving the stability and quality of GPT-5. The company plans to simplify the interface and make it clearer which model is being used. Users will also be able to manually enable the “Thinking” mode through the interface when this feature becomes available.
OpenAI acknowledges that the launch of GPT-5 was more challenging than expected. Due to the sharp increase in load, the company warns of a possible power shortage next week. Meanwhile, the team is optimizing the infrastructure and promises to inform about all changes in the service’s operation.
