API

In order to improve the system evaluation and leverage more performant models, we have chosen to compare the performance of the local model (Llama) with GPT-4o using an API. See the documentation.