As health-related digital platforms and their accessibility continue to expand for users, it is crucial to ensure accessible and comprehensible online information for audiences with limited literacy levels. This research focuses on the evaluation of large language models (LLMs) like ChatGPT and Llama-2 for text simplification tasks, specifically within the healthcare domain. We assessed these models using the Flesch-Kincaid Grade scale to determine their effectiveness in generating simplified responses. The results displayed a significant gap in providing adequate text simplification responses to users, leading us to develop an improved LLM model. Our approach involved instruction and fine-tuning an open-source LLM (Llama-2-7B) with a custom dataset containing the correct Flesch-Kincaid grade scale formula. A calculator function was implemented to guide our developed merged-model in generating responses appropriate to the user’s true reading grade level. Key findings indicate that the fine-tuned model consistently produced responses at or below the user’s reading grade level, significantly enhancing readability and accessibility of information, including the subject of health. This study contributes to the refinement of text simplification methodologies, aiming to improve health literacy and patient outcomes.