The presentwork is focused on the analysis of the static and free vibration of through-thickness functionally graded composite beams, where the material graded mixture joins a metallic phase, carbon nanotubes and a ceramic phase. In the beam plane containing its length and width a re-entrant honeycomb auxetic configuration is considered. The first-order shear deformation theory is used to perform the analyses, considering case-specific shear correction factors. Numerical applications are considered with a parametric study perspective to characterise the influence of different parameters in the mentioned mechanical responses. The predictions obtained are then analysed using a non-supervised machine learning method, the K-means clustering method, to identify groupings based on shared responses or characteristics, and subsequently to support decision-making for possible solutions.