Over the past few decades, ozone risk assessments for vegetation have been developed based on stomatal O₃ flux since this metric is more biologically meaningful than the traditional concentration-based approaches. However, uncertainty remains in the ability to simulate stomatal O₃ fluxes accurately. Here, we investigate stomatal O₃ fluxes simulated by six common air pollution deposition models across various land cover types worldwide. The Tropospheric Ozone Assessment Report (TOAR) database, a large collection of measurements worldwide, provides hourly O₃ concentration and meteorological data which are used to drive the models at 9 sites. The models estimated summertime O₃ deposition velocities of between 0.5–0.8 cm s^-1, mostly in agreement with the literature. Simulations of canopy conductance (G_st) showed differences between models that varied by land cover type with correlation coefficients of 0.75, 0.80 and 0.85 for forests, crops and grasslands. The model differences were determined by especially soil moisture and VPD depending upon the model constructs. Finally, the range of POD_y simulations at each site across models was most in agreement for crops (3 to 11 mmol O₃ m^-2) < forests (10 to 23 mmol O₃ m^-2) < grasslands (24 to 26 mmol O₃ m^-2). Nevertheless, ensemble model median response estimates gave results consistent with the literature in terms of those sites where O₃ damage is most likely to occur. Overall, this study is an important first step in developing and evaluating tools for broad-scale assessment of O₃ impact on vegetation within the framework of TOAR phase II.