This review explores the use of Large Language Models (LLMs) as virtual tutors in education, focusing on their applications in generating and evaluating educational materials. LLMs, based on Transformer architectures, have revolutionized various fields, including education, by enabling more flexible and autonomous learning environments. The review aims to provide a comprehensive overview of LLM-based solutions designed for educational purposes, particularly those involving students and teachers in their design or evaluation. The most popular models discussed are BERT, GPT-3, T5, and GPT-3.5. Common tasks performed by these models include question generation, answer grading, code correction, and explanation generation. The review also highlights the importance of prompt engineering and the need for transparency and fairness in LLMs, especially in educational settings. Despite the promising potential, the lack of reproducibility and ethical considerations remain significant challenges. The review concludes by suggesting future research directions, including the integration of LLMs into educational curricula and the exploration of innovative teaching practices.This review explores the use of Large Language Models (LLMs) as virtual tutors in education, focusing on their applications in generating and evaluating educational materials. LLMs, based on Transformer architectures, have revolutionized various fields, including education, by enabling more flexible and autonomous learning environments. The review aims to provide a comprehensive overview of LLM-based solutions designed for educational purposes, particularly those involving students and teachers in their design or evaluation. The most popular models discussed are BERT, GPT-3, T5, and GPT-3.5. Common tasks performed by these models include question generation, answer grading, code correction, and explanation generation. The review also highlights the importance of prompt engineering and the need for transparency and fairness in LLMs, especially in educational settings. Despite the promising potential, the lack of reproducibility and ethical considerations remain significant challenges. The review concludes by suggesting future research directions, including the integration of LLMs into educational curricula and the exploration of innovative teaching practices.