As building-integrated photovoltaics (BIPV) are a promising application of decentralized PV systems, it is important to facilitate the methods of their performance assessment. This requires efficient modelling and simulation environments that are able to use the strengths of different modelling tools for their designated purpose. This paper introduces a Python-based co-simulation framework to integrate the components of BIPV models that were developed in different tools using the Functional Mock-Up Interface (FMI). The presented BIPV model consists of two main sub-models; a coupled thermal-airflow model and an electrical model. These two models can be developed using different software tools that support the Functional Mock-Up Interface (FMI) approach, and then integrated into the proposed co-simulation framework. Subsequently, the framework is tested using a coupled thermal-airflow model developed in Dymola and two identical electrical models, one developed in Dymola and the other in Python. By exporting Functional Mock-Up Units (FMUs) from these tools and executing them within the introduced co-simulation environment in Python, the models were simulated simultaneously, while maintaining the same accuracy, this approach resulted in a reduction of more than 22% in simulation time compared to when both models were developed and exported using the same tool which is Dymola.