Integrating Large Language Models (LLMs) and generative artificial intelligence into educational practices will enhance text simplification tasks. Text simplification reduces the complexity of written content while maintaining its original meaning, making it more accessible to readers with cognitive disabilities, language learners, and young children. This paper examines the role of LLMs in text simplification, emphasizing their ability to rephrase complex sentences, correct grammatical errors, and replace difficult words with simpler alternatives. Benefits include improved comprehension, greater accessibility, increased engagement, and educational equity. Robust evaluation metrics are crucial to ensuring simplified texts’ quality and reliability, such as SARI, which measures simplification quality; BLEU, which evaluates fluency and accuracy; and ROUGE, which assesses content preservation, are discussed. This study evaluates four transformer-based models on the ASSET dataset: GPT-3.5 Turbo, GPT-4o, LLaMa 3, and Qwen 2. Results show different strengths and weaknesses, with GPT-4o as the best performer, achieving a BLEU score of 30.27 and a SARI score of 46.79. GPT-3.5 Turbo shows balanced performance, suitable for general educational use. With LLMs’ capabilities and understanding of their limitations, educators can enhance reading skills and inspire an inclusive learning environment. A comprehensive evaluation is necessary to thoroughly harness LLMs’ potential in educational text simplification.