Humans more and more live in urban areas and spend the majority of their time indoors, while buildings are becoming increasingly interconnected. Path planning in complex indoor environments is consequently becoming a societal challenge. This paper proposes a system, called BiMov, dedicated to automatically determining plan paths in complex buildings based on a digital mock-up (a BIM in IFC format). The originality of the research comes from the separation of the BIM model topological analysis from the semantic and geometric consideration of the customer of the path planning service. The process consists in exploiting the semantic, geometric, and topological features of the constituents of a BIM to generate navigation graphs that take the profile of MOoP (Mobile Object or Person) and the operational accessibility status of spaces and transitions (such as doors or stairs) into account. The ultimate aim is to determine the optimal path according to given criteria. The system is based on four data models: (1) a building model derived from the original BIM that represents and structures the building features relevant to indoor mobility; (2) a MOoP model representing its bulk size, ability to move horizontally and vertically, and various authorisations; (3) a calendar model representing the conjectural accessibility status of spaces and transitions; (4) a navigation graph model proposing three levels of detail. The Macro level represents a simple graph of connectivity between neighbouring interior spaces and is intended to help architects verify their design in terms of accessibility. The External level is used to connect accessible spaces via their common horizontal or vertical transitions. This level is intended for MOoPs that do not require a detailed description of the path. The Internal level integrates meshing of each space: a 2D mesh for planar mobility or a 3D mesh for drones. This level is intended for MOoP such as heavy object handlers or autonomous mobile vehicles that need to validate a reliable path within spaces containing furniture, hazardous machinery, or bulky equipment. We developed a prototype software application that illustrates our approach for different path planning scenarios in a BIM model generated outside this project.