Evaluation: Measuring our Agents capabilities with Gaia

When working with the non-deterministic nature of LLMs (the same prompt will return different responses each time, due to the probabilistic nature of LLMs) traditional means of software testing are no