OpenAI releases GPT-4.5 to improve the breadth of knowledge and the naturalness of interaction

PANews reported on February 28 that according to OpenAI announcement, GPT-4.5 was officially released, and the first batch has been opened to ChatGPT Pro users, covering Plus and Team users next week, and then promoted to Enterprise and Edu users. This model has improved compared with previous generations in terms of knowledge breadth, understanding ability and user interaction experience, and has reduced the phenomenon of "illusion".

GPT-4.5 adopts larger scale unsupervised learning to enhance pattern recognition, logical connection and creative generation capabilities, but does not have the ability to reason. Its improvements include stronger emotional intelligence (EQ), more natural conversational experiences, and enhancements in writing, programming, and automation of complex tasks. In addition, this model supports file and image input, but does not support voice mode, video and screen sharing for the time being. OpenAI said that in the future, it will optimize the user experience to make AI more intuitive and easy to use.