The building sector’s significant contribution to global energy consumption and greenhouse gas emissions necessitates more intelligent building design and management, leading to the growing importance of Building Digital Twins. This study focuses on functional, simulation-based digital twins that incorporate physics-based models for dynamic building performance simulation. The research investigates the interoperability of automatically generated Modelica models using TEASER (Tool for Energy Analysis and Simulation for Efficient Retrofit) and the AixLib library within the Modelon Impact simulation environment. TEASER was employed to generate reduced-order thermal building models from archetype data, which were then simulated using Modelon Impact, requiring a hybrid approach to address software compatibility issues. A comprehensive case study of a multi-level residential facility was conducted, evaluating thermal comfort and energy consumption at different levels of aggregation. The results demonstrate the feasibility of this integrated workflow for creating dynamic energy models, highlighting the trade-offs between modelling granularity and accuracy. The study contributes to more flexible and robust digital twin implementations for enhanced building performance analysis and operational decision-making.