OpenAI responds to complaints that GPT-5 feels less friendly

The new GPT-5 model has faced criticism from users who say it produces overly formal replies.

The familiar friendly tone many had grown accustomed to seems to have disappeared. OpenAI acknowledged the feedback and promised to make the chatbot warmer and more approachable:

You'll notice small, genuine touches like "Good question" or "Great start," not flattery. Internal tests show no rise in sycophancy compared to the previous GPT-5 personality.

Users began voicing concerns almost immediately after launch, claiming GPT-5 showed "regression disguised as improvement." In response, the company later allowed users to switch between GPT-5 modes and restored access to several older models in ChatGPT.

OpenAI brings back model selection in ChatGPT — including some older versions

OpenAI has reinstated model selection in ChatGPT, letting Plus users choose between GPT-5 modes and several older models.

Durov's CodeIgor Savkin