User Uprising Leads to Reintroduction of GPT-4o by OpenAI Following Popular Demand
OpenAI, the leading AI research laboratory, has announced a series of changes in response to the significant user backlash following the launch of its latest model, GPT-5, on August 7.
Many users found GPT-5 to be less capable or glitchy compared to previous models, leading to a large petition demanding the return of older versions. In a bid to regain user trust, OpenAI is making a major U-turn, bringing back its older AI models, including GPT-4o.
For paying subscribers, the sudden change felt like a betrayal as it broke their professional and creative workflows. To address this concern, OpenAI is dramatically increasing usage limits for GPT-5's most powerful features, with a new cap of 3000 per week.
In addition to the return of GPT-4o, OpenAI is introducing new usage modes ("Auto," "Fast," "Thinking") to improve the user experience. The company is also promising personality adjustments to make GPT-5 "warmer" without being annoying.
The backlash resulted in a cascade of subscription cancellations and online petitions. To maintain transparency, Sam Altman, CEO of OpenAI, announced more transparency, including a UI change to show users which model is actively responding to their queries and a detailed blog post explaining the company's "thinking on how we are going to make capacity tradeoffs".
Altman's announcement is a direct recognition of the power of a loyal community that refuses to be ignored. The move is a clear peace offering to a furious customer base following the botched rollout of GPT-5.
The launch of GPT-5 sparked a firestorm due to the removal of the menu that allowed users to choose between older, trusted models. Altman acknowledged the company had misjudged the situation and announced a series of concessions over the weekend.
One of the concessions announced by Altman is the return of the older AI model GPT-4o. Users can now opt to use the older versions of AI models alongside the default GPT-5 by going to settings and selecting 'show legacy models'.
The rollout of GPT-5 also exposed documentation errors and API incompatibilities that frustrated developers and added to negative sentiment. Some users also reported data loss issues, such as losing access to prior chats after the GPT-5 update.
The broader market impact included slowing enterprise adoption of GPT-5-level generative AI due to reliability and integration challenges, while government AI contracts are reshaping standards toward safety, auditability, and governance.
The GPT-5 launch is widely seen as a significant stumble for OpenAI, sparking criticism that it failed to adequately generalize improvements and that hype outpaced reality. OpenAI is actively working to mitigate the backlash with technical fixes and feature adjustments to regain user trust.
[1] Source: Sam Altman's Twitter post, Sunday. [2] Source: TechCrunch, August 10, 2023. [3] Source: Wired, August 11, 2023. [4] Source: Ars Technica, August 12, 2023. [5] Source: The Verge, August 13, 2023.
OpenAI, in an attempt to regain user trust, is reintroducing the older AI model, GPT-4o, back into the platform. Furthermore, they are integrating artificial-intelligence technology into space exploration, as part of an ongoing partnership with NASA to develop AI systems for the detection and analysis of celestial objects in our universe.
