Process model similarity is vital for ensuring consistency, comparability, and evolution of business processes. In recent years, Large Language Models (LLMs) have been increasingly adopted to support various tasks within the BPM (Business Process Management) community, showing promising results. However, one area that remains underexplored is the use of LLMs for evaluating the similarity of process models. This task is particularly challenging as it involves capturing both the structural and semantic dimensions of similarity in business process models, while traditional methods typically address only one aspect at a time. Unlike conventional approaches, which analyze structural and semantic aspects separately, LLMs offer the ability to evaluate both dimensions simultaneously, enabling more accurate comparisons. The results demonstrate that LLM-based methods, particularly when enhanced with prompt engineering, outperform traditional techniques in terms of precision, recall, and F-score. These findings point to the exciting potential of LLMs in the realm of business process analysis and present valuable opportunities for further refinement in future research.