Fig. 1: Large language models (LLMs) evaluated in this study. | BDJ Open