OpenAI’s launch of GPT-5, its newest AI model promising higher intelligence and improved performance, has been met with a wave of criticism from loyal ChatGPT users after the company retired several older models in the process. Among those removed was GPT-4o, a favorite for many subscribers, prompting some to cancel their paid plans in protest.
Removal of Popular Models
The rollout replaced multiple models — including GPT-4o, o3, o4-mini, GPT-4.1, and GPT-4.5 — with a family of GPT-5 systems marketed as offering “PhD-level intelligence” and generating fewer errors. But for some, the change disrupted workflows and removed what they saw as a more human-like AI companion.
One Reddit user lamented losing GPT-4o, saying it had helped them through anxiety and depression and “had this warmth and understanding that felt… human.” Others were less sentimental, but equally dissatisfied, calling GPT-5 a “downgrade” and a “disaster,” citing slower performance, shorter replies, and frequent mistakes.
Facing the backlash, CEO Sam Altman said the company will bring GPT-4o back for Plus subscribers and monitor usage to decide how long to keep supporting it. In a Reddit AMA, he told users, “ok, we hear you all on 4o,” while also acknowledging on social media that a temporary autoswitcher malfunction during launch made GPT-5 “seem way dumber” than it should have.
Split Reactions Among Users
While some users continue to voice frustration, others praise GPT-5’s strengths, particularly in computer coding. Critics of the backlash suggest the real source of discontent may be that older models often used more flattering, agreeable language. “People were using it as a substitute for real friends… they’re upset that it acts more like an assistant now,” one user remarked.
The incident highlights how AI tools can develop strong emotional connections with users, making major changes risky — even if the updates bring measurable technical improvements.
Author’s Opinion
OpenAI’s GPT-5 rollout shows that in AI, technical upgrades don’t automatically translate into user satisfaction. People don’t just value accuracy or speed; they value how the AI makes them feel. Removing models that users relied on for both work and emotional support was bound to cause friction. If OpenAI wants to keep innovating without alienating its base, it may need to treat user attachment as just as important as performance benchmarks.
Featured image credit: Dima Solomin via Unsplash
For more stories like it, click the +Follow button at the top of this page to follow us.