The predictivity of models utilized in drug development remains a persistent challenge, hindering the successful translation of preclinical findings to clinical outcomes. This challenge can be resolved into three distinct translational hurdles: species translation, where animal models often inadequately represent human responses; complexity translation, as in vitro models frequently fail to capture the full intricacies of in vivo systems; and human diversity translation, where the controlled laboratory conditions diverge greatly from the heterogeneous clinical environment. In this session, we will explore innovative strategies aimed at addressing these translational problems and enhancing the predictivity of drug development and safety models. We will feature talks including advanced and physiological human-cell based models, big data integration and multi-scale computational models. This meeting is endorsed by the sector plans of pharmaceutical sciences and of medicine.