This study presents the development, simplification, and validation of building energy models (BEMs) for a large educational facility designed as zero energy building (ZEB) with renewable energy resources (RES) hybrid RES systems, including district heating, geothermal heat pumps, PV panels and solar thermal. A detailed reference model comprising 197 zones was developed using IDA indoor climate and energy (ICE) software and subsequently simplified into three reduced-order models to improve computational efficiency. The second reduced model (22 zones) integrates an early stage building optimization (ESPO) plant configuration that enables coordinated charging and discharging of a thermal storage tank by three independent sources under separate control strategies. Model validation was conducted against experimental energy measurements from 20 power and heat meters. Reduced Model 2 demonstrated the highest accuracy, achieving a mean percentage error (MPE) of 12.65 % and a mean absolute percentage error (MAPE) of 14.03 %, while reducing computational time from 45.06 to just 1.50 h. Thermal comfort analysis confirmed that operative temperatures exceeded 27 °C in less than 7 % of hours in the worst zone. The results highlight that reduced-order models, when equipped with an advanced and integrated plant representation, can offer reliable accuracy with significant time savings. Such models are well suited for early design evaluation, control strategy assessment, and performance monitoring in complex, energy-efficient buildings.