LLMs Still Can't Avoid Instanceof: An Investigation Into GPT-3.5, GPT-4, and Bard's Capacity to Handle Object-Oriented Programming Assignments

Large Language Models (LLMs) struggle with object-oriented programming, emphasizing the importance of code quality and supervision in educational settings.