This paper presents a study on the application of ChatGPT 4.0 in developing a numerical model for the performance analysis of a turbofan engine. The modeling process began with general queries regarding numerical approaches to engine simulation. The initial model proposed by ChatGPT appeared plausible but contained significant conceptual errors. Through iterative dialogue and refinement, these errors were gradually identified and corrected, ultimately resulting in a valid engine model. This intermediate model included two rotating components (fan and core spool) and assumed an ideal gas with distinct thermodynamic properties in the cold and hot sections of the engine. Based on this model, ChatGPT successfully generated numerical code for implementation in the MATLAB environment, handling this task with high accuracy and flexibility. Further efforts focused on extending the model to include air extraction for turbine cooling, internal engine bleeds, and the application of a semi-perfect gas model to describe the working fluid more realistically. In these more advanced areas, ChatGPT’s performance declined significantly. Despite prompting and corrective guidance, it was unable to produce a fully functional and physically accurate implementation of the enhanced model. The study concludes that while ChatGPT demonstrates strong capabilities in translating well-defined physical models into numerical code, especially within MATLAB, it remains unreliable in constructing or modifying complex thermodynamic models without significant user oversight. Nonetheless, its use can significantly accelerate the implementation phase of numerical engine modeling when guided by an experienced user.