One of the avenues to integrate more renewable energies is to contract flexibility through managing the energy use of building stocks, through incentives and penalties. There are different incentive-aware control strategies utilized for districts to tap that flexibility, such as Rule Based Control (RBC) and Model Predictive Control (MPC). MPC as one of the most potent control strategies is capable of optimizing the cost of the households over a certain time horizon, and subsequently provide flexibility through receiving an incentive signal, such as the price of electricity, alongside using an optimizer and a predictive model of the building. The provided flexibility characteristics however, may be impacted by the type and quality of the predictive model of the building, as it captures the physics of the building. One of the methods to investigate the provided flexibility is the concept of Flexibility Function (FF), which has six main Key Performance Indicators (KPIs). and the FF could characterize the flexibility dynamics in a quantitative manner for the future contractions. This concept has mainly been applied on individual buildings so far. In this study, it will be applied on district level. To do so, a dummy district is generated in Dymola using the Modelica language through the IDEAS library. Then a dataset is generated for each of the buildings and two different modelling techniques are used to develop the predictive models for each of them. Afterwards, an MPC is casted for each building and a price signal (impulse signal) is broadcasted to them to realize the response of the district. Finally, using the concept of FF and its KPIs, it is investigated how a certain modelling technique for the predictive model is impacting the overall flexibility of the entire district.