In the information era, the common practice is to build 3D numerical models of objects in 3 dimensions to map and predict the nature and geometries of subsurface objects at different scales. These models aim at integrating all available information and knowledge to improve consistency and manage subsurface uncertainty. Day-to-day challenges are associated to 3D geomodelling, e.g. the efficient use of databases to produce geological models, the relevant modelling and complex geological concepts and/or architectures, the use of huge amount of data (point clouds, geophysics...), and the versioning or/and reactive update of models.The session will focus on leading-edge subjects of current research and developments to improve the accuracy, reliability, lifecycle management and usability of 3D geomodels as both as knowledge integration per se and as support for applied computations ranging from resource estimation to numerical simulation of physical processes.