'We hear you': Sam Altman restores GPT-4o for plus subscribers after GPT-5 backlash
During a Reddit Ask Me Anything (AMA) session, several users directly told OpenAI CEO Sam Altman that GPT-5 felt “dumber” than its predecessor and pleaded for the return of GPT-4o.
ADVERTISEMENT
Earlier this week, ChatGPT maker OpenAI officially unveiled GPT-5, touting it as its most advanced and capable large language model to date.
The new AI is now available to more than 700 million users worldwide. However, the rollout has triggered a wave of frustration, with both free and paid ChatGPT customers complaining about losing access to the older GPT-4o model - a version many had integrated deeply into their workflows.
What exactly happened?
During a Reddit Ask Me Anything (AMA) session, several users directly told OpenAI CEO Sam Altman that GPT-5 felt "dumber" than its predecessor and pleaded for the return of GPT-4o.
Responding in the same thread, Altman confirmed that the company had "heard the feedback" and will allow Plus users to continue using GPT-4o, while monitoring usage to decide how long to support it.
The criticism isn't just about performance. Many users were caught off guard by OpenAI's abrupt removal of legacy models, saying they had no prior warning that their preferred AI would be phased out.
A key new feature of GPT-5 - a real-time router that automatically decides which model to use for a given task - has also come under scrutiny.
Altman admitted the router "wasn't working as intended," causing the model to seem less intelligent, but assured users that it "will seem smarter starting today." He also promised greater transparency around which model is answering a query.
Currently, ChatGPT Pro, Team, and Enterprise customers can revert to GPT-4o by enabling the “show legacy models” option in settings. Free users, however, have no way to access older models. When asked about bringing back GPT-4.1, Altman said it would only happen if there was “enough demand” from Plus subscribers.
Closing the AMA, Altman pledged that OpenAI will “continue to work to get things stable” and “keep listening to feedback” as GPT-5 adoption grows.
