This paper presents the first known study to characterize the performance of passive chilled beams operating in the field. Extensive field measurements were conducted in an open-plan office space equipped with multiple passive chilled beam units. The experiments included variable water mass flow rates, supply air flow rates to the space, and air-water temperature differences. The experimental results were used to (i) test the validity of a simple model described in Standard EN 14518 and identify weaknesses and (ii) to develop a new regression model that can be used to characterize the performance of installed passive chilled beams under different operating conditions. The EN 14518 model could not capture the effects of water mass flow rate, supply air flow rate, and potential interaction effects between multiple passive chilled beams operating in the same space. A new regression model, based on experimental results, was therefore developed to predict the cooling capacity of passive chilled beams under variable operating conditions (water mass flow rate, supply air flow and air-water temperature difference). The new model uses normalized supply water temperature and water mass flow rate as the main correlating parameters. Supply air flow rate for the ventilation system did not significantly affect the cooling capacity for the studied case. The new model showed good agreement with experimental measurements and can be used in building energy modeling tools to predict the performance of passive chilled beams installed in similar indoor spaces.