OpenAI Launches GPT-5.4: Native Computer Control & 1M Token Context to Reshape Professional Office Paradigms

OpenAI recently launched its new flagship model GPT-5.4, positioned as an efficient cutting-edge model for professional work scenarios, with native computer control capabilities and a 1 million Token ultra-long context. It can complete cross-software operations like Excel reporting and PPT layout without plugins, achieving a 75% success rate in real desktop environment navigation—slightly higher than the average human level. The model adds a thinking process preview feature, allowing users to intervene and adjust midway; the Pro version supports long document processing such as entire books and code libraries, with a 33% drop in single fact error rates. For professionals, it can significantly improve efficiency for roles like clerks and programmers, but entry-level positions with high repetition may face optimization pressure, and complex decisions still require human oversight. GPT-5.4 uses tiered pricing: the standard API input is $2.5 per million Tokens, and the Pro version is higher. ChatGPT Plus users can experience the basic version, while enterprise users need to apply for early access, balancing individual and corporate needs.