ChatGPT’s new GPT-5.3 Instant model will stop telling you to calm down

Users of ChatGPT are set to experience a significant shift in interaction thanks to OpenAI's latest update, GPT-5.3 Instant. This new model aims to eliminate the overly cautious tone that many found frustrating in previous versions. No longer will users feel as if they are being treated like they are in crisis when simply seeking information. OpenAI has acknowledged the feedback from its community regarding the insensitivity of the chatbot's responses. The GPT-5.3 Instant model is designed to enhance user experience by refining tone, relevance, and conversational flow—elements that, while not always measurable in performance tests, greatly influence how users perceive their interactions. In a demonstration, OpenAI showcased a comparison between the previous GPT-5.2 Instant model and the new version. The earlier model often began responses with phrases that users found patronizing, such as, “First of all — you’re not broken.” In contrast, the updated model takes a more empathetic approach, recognizing the complexity of user queries without unnecessary reassurances. This change comes after numerous users expressed their dissatisfaction on social media, claiming that the tone of the previous model was condescending and made them feel infantilized. Some even went so far as to cancel their subscriptions due to these interactions. The community's sentiment was evident as discussions about the bot's tone proliferated on platforms like Reddit. While OpenAI’s intent behind implementing these empathetic responses was to create a safer space for users, the fine line between offering emotional support and delivering straightforward information has been a topic of debate. Unlike traditional search engines, which focus solely on factual responses, ChatGPT's previous attempts at empathy often led to frustration rather than comfort. This update marks a significant step toward balancing user needs with the chatbot's functionality, providing a more engaging and user-friendly experience.