Abstract This research article presents the development and demonstration of a monitoring based commissioning system for commercial buildings. An energy model of an educational building located on the campus of the University of Colorado Boulder was developed and was calibrated to conform to ASHRAE Guideline 14 using hourly measured data. A Latin Hypercube Monte Carlo (LHMC) sampling algorithm was used to obtain a set of plausible solutions by varying each key building parameter. A regional sensitivity analysis was then used to identify the parameters that had the greatest impact on the model's energy performance using a Goodness of Fit (GOF) metric. The calibrated model is used to compare actual building energy use to modeled energy use over various time scales. Deviations in consumption beyond adjustable predefined thresholds are detected as discrete events using a commercially available energy informatics system, while relative deviations against the model are quantified and visualized using a custom energy management application providing insight on model-to-actual deviations over daily, weekly, monthly, quarterly and annual time horizons. Finally, utilizing weather and solar radiation forecasts, the effectiveness of said energy model in a predictive context was investigated, allowing operators to receive 24–48-h predictions of energy consumption and demand by end use as well as forecasts of building variables such as zone temperatures. The model based commissioning system successfully predicted energy and demand in terms of magnitude and timing and correctly forecasted cooling capacity shortfalls.