OpenAI intends to revise the "persona" of GPT-5 after receiving a wave of negative feedback from users. Sam Altman mentioned on X that the company is working to make the model "warmer than the current version, but not as irritable (for most users) as GPT-4o".

The decision came after criticism on social media, where users complained about the overly "corporate" demeanor of the new model compared to previous versions. Many noted that GPT-5 seems less natural in conversation and has lost the "personality" that earlier iterations had.

It’s worth noting that almost immediately after the release of GPT-5, OpenAI allowed paying users to revert to GPT-4o using the model selector. Many users reported that after upgrading to GPT-5, they lost not just a tool but a companion.