OpenAI to Bring Back GPT-4o After GPT-5 Backlash
Readers help support Windows Report. We may get a commission if you buy through our links.
Read our disclosure page to find out how can you help Windows Report sustain the editorial team. Read more
OpenAI launched GPT-5 earlier this week, calling it its most advanced model yet. The release put the new AI in the hands of more than 700 million people worldwide.
What should have been good news for OpenAI turned into a nightmare as it faced backlash from both free and paid ChatGPT users. The problem many discovered that GPT-4o, a model they had built workflows around, had disappeared without warning.
During a Reddit AMA meant to celebrate the launch, CEO Sam Altman faced a wave of complaints. Several users said GPT-5 felt “dumber” than its predecessor and demanded GPT-4o’s return.
Altman replied that OpenAI had “heard the feedback” and would let Plus subscribers continue using GPT-4o, while tracking usage to decide how long it stays available. The rollout wasn’t just criticized for performance.
The removal of legacy models without notice left some feeling quite frustrated. At the center of the frustration is GPT-5’s new “real-time router,” a system that automatically picks the model for each task.
Altman admitted the router “wasn’t working as intended,” making GPT-5 appear less capable. He assured users it “will seem smarter” and promised more transparency about which model is responding.
For now, only Pro, Team, and Enterprise customers can toggle GPT-4o back on via the “show legacy models” setting. Free users don’t have that option. As for GPT-4.1, Altman said it would only return if Plus subscribers showed “enough demand.”
Closing the AMA, Altman pledged to “continue to work to get things stable” and to “keep listening to feedback” as GPT-5 rolls out.
