To ensure that ML and AI models can maintain solid and consistent performance, even when faced with variable conditions, data noise, or environmental disturbances, it is necessary to evaluate them under different operating conditions. VRAIN researchers have extensive experience in analyzing the robustness of models, understanding it as the ability to generalize well to different data sets and effectively handle unexpected situations. At VRAIN, we have developed specific methodologies for evaluating the robustness of models by measuring the impact that data complexity and noise levels have on the models.