GNI BIM Dataset
224 anonymized IFC building models from student coursework at the TUM Georg Nemetschek Institute – AI for the Built World / Chair of Computing in Civil and Building Engineering, Technical University of Munich, released with the authors' consent under CC BY 4.0.
224 anonymized IFC building models from student coursework at the TUM Georg Nemetschek Institute – AI for the Built World / Chair of Computing in Civil and Building Engineering, Technical University of Munich, released with the authors’ consent under CC BY 4.0.
2025_BIMfundamentals/ — 208 single-discipline models from the BIM Fundamentals master course (Summerterm 2025). Each model was created individually by a student in a commercial BIM tool.
2026_BIMprojects/ — 16 models from 9 teams in the BIM Project master course (Winterterm 2025/26). 7 of 9 teams produced paired architectural + structural models of the same building.
Identifying information (author names, organizations, addresses, project metadata) was removed from all IFC files. Anonymization scripts are available at https://github.com/ZijianWang-ZW/GNI-BIM-Dataset.
Across 223 parsed models, the dataset contains 290,877 element instances spanning 42 IfcElement subclasses (walls, slabs, beams, columns, doors, windows, structural framing, MEP, furnishings, and more). Models use IFC2x3 and IFC4 schemas.
A CSV file with links to 35 other online IFC model collections is included.
📤 Share this page
Found this useful? Share it with your network.
Files are hosted on the source repository. Click download to access the full dataset.