In an era of rapidly expanding digital archives, effective text summarization tools are crucial for educators and learners. This study presents an advanced text summarization system specifically designed for Arabic textbooks, leveraging state-of-the-art natural language processing (NLP) models. We trained and evaluated four models - AraBART, mBART50, AraT5, and MT5 - on biology textbook content for 11th and 12th grades in the Palestinian curriculum using the ROUGE metric to compare their performance. Our results show that AraBART and mBART50 outperformed the other models in terms of ROUGE-1 (score: 0.24), ROUGE-2 (score: 0.118–0.139) and ROUGE-L (score: 0.24), indicating their effectiveness in summarizing complex biological information while preserving essential details. To further validate the performance of these models, we conducted a subjective evaluation by an educational expert who assessed the summaries for their accuracy, relevance, and overall quality, with AraBART ranking highest due to its ability to capture key concepts and reduce noise in the original text. The study’s findings provide valuable insights into the performance of these models and contribute to the development of effective tools for understanding and generating Arabic text.