GPT-4o Returns to ChatGPT Following User Outcry
After a significant backlash from the ChatGPT community, OpenAI has reinstated the GPT-4o model within its platform. This decision comes shortly after the model was removed during the rollout of GPT-5, sparking widespread dissatisfaction among users who favored GPT-4o’s unique capabilities. Now, GPT-4o is once again accessible by default to all paying subscribers, including those with ChatGPT Plus accounts, through the model selection menu.
User Feedback Drives OpenAI’s Reversal
OpenAI CEO Sam Altman acknowledged that the company underestimated how much users valued GPT-4o’s features. In an effort to streamline the AI offerings, OpenAI initially eliminated all previous versions from ChatGPT, which triggered the most intense user protest in the platform’s history. A Reddit discussion titled “GPT-5 is horrible” quickly amassed over 2,000 comments within days, highlighting the community’s frustration and demand for the return of earlier models.
Enhancements Alongside GPT-4o’s Return
In addition to restoring GPT-4o, OpenAI has implemented several improvements to enhance user experience. The weekly message limit for GPT-5’s “Thinking” mode has been dramatically increased from 200 to 3,001 messages, accommodating heavier usage. Once this threshold is reached, users can switch to the new “GPT-5 Thinking Mini” option, which offers additional capacity. Furthermore, OpenAI introduced three routing modes-“Auto,” “Fast,” and “Thinking”-allowing users to select the GPT-5 variant that best suits their needs for each query.
Screenshot of ChatGPT’s model selector interface as of August 13, 2020.
Image credit: Benj Edwards
Exclusive Features for Pro Subscribers
Altman also revealed plans to expand the model lineup for Pro users, who pay $200 monthly. These subscribers will gain access to additional models such as o3, GPT-5 Thinking Mini, 4.1, and 4.1 through a new “Show additional models” toggle in the web settings. However, GPT-4.5 will remain exclusive to Pro members due to its substantial GPU resource requirements, reflecting the high operational costs associated with running advanced AI models.
Looking Ahead: Balancing Innovation and User Preferences
This episode underscores the challenge OpenAI faces in balancing cutting-edge AI development with the preferences of its diverse user base. By reintroducing GPT-4o and expanding customization options, OpenAI demonstrates responsiveness to community feedback while continuing to push forward with GPT-5 innovations. As AI adoption grows, such adaptive strategies will be crucial in maintaining user satisfaction and engagement.
