In our examination of the IEP evaluation’s failure instances, we sought to discover the things limiting LLM general performance. Presented the pronounced disparity between open-supply models and GPT models, with a few failing to generate coherent responses constantly, our Assessment centered on the GPT-4 model, essentially the most advanced model