Sam Altman vows to make GPT-5 ‘warmer’ after users complain ChatGPT has lost its charm

After users expressed dissatisfaction with the new AI model’s lack of humor, inventiveness, and relatability, OpenAI CEO Sam Altman pledged on Saturday that the company would endeavor to make GPT-5 “warmer” in talks. Altman added that even while the new model performs better in many aspects, his team had understated the significance of some of the features that consumers found appealing about GPT-4o.

OpenAI declared shortly after the release of GPT-5 that it was discontinuing all of its previous models, including GPT-4o, GPT-4.1, GPT-4.5, GPT-4.1-mini, o4-mini, o4-mini-high, o3, and o3-pro. There was a great deal of outcry when GPT-5 began to roll out to users and they were no longer able to talk with the previous iterations of the chatbot.

According to OpenAI’s launch blog, GPT-5 is faster and more accurate than all of its previous models, and it has fewer instances of sycophancy (being excessively agreeable) and hallucination (making things up). Significant progress was also reported by the AI startup in a number of tasks, such as writing, thinking, coding, answering health-related queries, and multimodal skills.

Social media responses, however, indicate that GPT-5 may occasionally be slower than earlier models and provide shorter responses with less nuanced emotional expression. After the backlash on social media, Altman took to X to share the changes that will be coming to GPT-5 in the coming days and weeks.

“We for sure underestimated how much some of the things that people like in GPT-4o matter to them, even if GPT-5 performs better in most ways.” Altman wrote. “We are going to focus on finishing the GPT-5 rollout and getting things stable and then we are going to focus on some changes to GPT-5 to make it warmer. Really good per-users customization will take longer.” he added .

Altman pointed out that customers’ perceptions of GPT-4o’s advantages over GPT-5 varies greatly, and OpenAI will eventually try to provide users more customization options. According to Altman, for example, some users like emojis in ChatGPT responses, while others prefer “cold logic,” and still others prefer warmth and a distinct form of emotional intelligence.