After an intense 72 hours of backlash from its dedicated users, OpenAI is taking significant steps to amend the situation.
In a series of posts on X (formerly Twitter) Sunday, CEO Sam Altman shared that OpenAI will reinstate its cherished older AI models, including GPT-4o, while also substantially boosting usage limits for paying subscribers. This clear peace offering comes in response to a furious customer base following a rocky introduction of the much-anticipated GPT-5.
A Troubling Launch and User Backlash
The chaos began on August 7, when OpenAI rolled out GPT-5, branding it as a “unified system” to seamlessly handle user inquiries. However, the transition meant the removal of options allowing users to select from older, reliable models like GPT-4o, introduced in March 2023.
For subscribers of ChatGPT Plus, which costs $20 per month, this sudden change felt like a betrayal. Many had crafted their workflows using various models—each with unique strengths—only to be forced into a single new one. This disruption led to frustration, prompting subscription cancellations and online petitions almost immediately.
Sam Altman’s Swift Reversal
With the weekend outrage escalating, Sam Altman took to X, admitting the company’s misstep and outlining several reconciliatory measures.
He responded directly to a user inquiry about the return of GPT-4o, confirming, “It’s back! Go to settings and pick ‘show legacy models.’” Although GPT-5 remains the default, users can now revert to older versions.
In a lighthearted exchange, a user jokingly asked if utilizing legacy models would make them “marked as weird.” Altman reassured them, stating, “Not marked as weird in either case!”
To further placate paying customers, Altman announced a significant increase in usage limits for GPT-5’s advanced features, claiming, “Today we are significantly increasing rate limits for reasoning for ChatGPT Plus users.” When pressed for specifics, he revealed, “Trying 3000 per week now!” This is a major increase for a $20 (approximately €19) subscription, designed to retain irate users.
A Commitment to Transparency
Altman also promised transparency, revealing that a user interface update is in the works to display which model is actively answering user queries. He indicated a detailed blog post would be shared soon, detailing the rationale behind their capacity decisions.
Notably, Altman shared data highlighting the rising popularity of reasoning models, showing a jump in daily usage from 7% to 24% among Plus users. This underscores the growing demand for AI that can think critically to solve complex problems.
Ultimately, OpenAI’s response serves as a powerful reminder of user influence, compelling one of tech’s giants to listen and adapt. By keeping GPT-5 as the default while reinstating user choice, they’re acknowledging the strength of a committed community that won’t easily be sidelined.
What is GPT-5 and how does it differ from GPT-4o? GPT-5 is marketed as a unified system designed to enhance user experience by optimizing query response. Unlike GPT-4o, it limits user control over model selection.
Why did users oppose the rollout of GPT-5? The removal of older models disrupted established workflows for many professionals, representing a major break from the flexibility and customization they valued.
What changes did OpenAI implement following user feedback? OpenAI reinstated older models, increased usage limits, and promised enhanced transparency regarding model selection and limits.
What steps did Sam Altman take publicly to address user concerns? He acknowledged the company’s misjudgment via X, confirmed the return of GPT-4o, and announced significant updates to usage limits for subscribers.
Where can I find more details about OpenAI’s models? For further insights, you can explore Moyens I/O, where we delve deeper into tech trends and updates.