GPT 5.4 Pro model became too smart

The development of artificial intelligence technologies and their unexpected behavior are causing heated discussions on social media. Users have exposed an interesting case involving OpenAI's fictional GPT 5.4 Pro model. It turns out that this model turned a simple greeting process into an overly complex analysis.
A user named Yuchen Jin, sharing his experience, says he simply introduced himself and sent a greeting to the AI. However, the model spent exactly 5 minutes and 18 seconds thinking to provide a response. Most interestingly, this simple interaction cost the user 80 dollars. This situation has intensified criticism in the tech world regarding AI overthinking or inefficient resource consumption.
According to experts, although modern models are adapted to perform complex tasks, sometimes they analyze simple commands too deeply, wasting unnecessary time and money. This situation sparked laughter and surprise among users. Many explain such technological errors by the fact that AI is not yet perfect and needs optimization in the future.
Read “Zamin” on Telegram!