OpenAI announced earlier this week that its newest language model, GPT-5, was available to all ChatGPT users. However, this transition, which completely disabled existing models GPT-4o and GPT-3, sparked backlash. Paid Plus subscribers, in particular, are complaining about the limitations imposed by the new model.
ChatGPT Subscribers Dissatisfied with Limits
Plus subscribers, ChatGPT’s most loyal user base, are disappointed with the new limits imposed on GPT-5. Users can only send 80 messages every three hours. Once this limit is reached, the system automatically switches to the lower-capacity GPT-5 mini model. Subscribers claim they have unlimited access to previous models like GPT-4o. They argue that this eliminates the benefits of a paid membership.
OpenAI, not ignoring the barrage of user criticism, has announced an urgent action plan based on the feedback it has collected over the past two days.
Here are the steps taken:
Limits Doubled: Weekend message limits have been doubled for users on the Plus and Team plans.
New “Mini” Models Coming: Mini versions of “GPT-5 Thinking” and “GPT-5 Pro” will be available next week, which will be activated when the main model’s limit is reached.
Model Selection Becomes Flexible: Users will now be able to access “GPT-5 Thinking” and “GPT-5 Pro” directly from the main model selector.
The Legend Returns: GPT-4o is Back: One of the most requested changes has been made. Users will be able to reuse their beloved GPT-4o model by selecting the “Show older models” option in the interface.
OpenAI CEO Sam Altman made the announcement during an Ask Me Anything (AMA) on Reddit. Altman stated that they will be closely monitoring how long users will continue to use GPT-4o. He added that support will be shaped based on this data.
Another highlight from the event was a new subscription plan. Altman hinted that they are working on a new middleware plan between the $20 per month Plus and $200 per month Pro packages. This plan will offer greater flexibility.
They also acknowledged a technical issue with the model switcher on the first day of the GPT-5 deployment, which resulted in insufficient responses to some queries.
Michelle Pokrass from the OpenAI team stated that they would like to offer longer content support for API users, up to 1 million tokens. However, she explained that this is not currently a priority due to current high GPU demand and computational constraints.
It appears OpenAI is trying to balance listening to its massive user base of 700 million. These steps are crucial to demonstrating the company’s commitment to community feedback and its flexibility.
