The use of artificial intelligence (AI) based on Large Language Models (LLMs), such as GPT-4, GPT-3.5, and Gemini, is becoming more prevalent in software development, particularly in identifying and refactoring code to adhere to SOLID design principles. This study analyzes the effectiveness and accuracy of AI in detecting and refactoring code to follow these principles. We used code samples from second-semester Informatics Engineering students at UIN Malang as the dataset. The analysis was performed by comparing the SOLID principle detection results from GPT-4, GPT-3.5, and Gemini. The findings show that Gemini is the most accurate in identifying and applying SOLID principles, achieving a perfect score compared to GPT-4 and GPT-3.5, which had lower accuracy levels.