Three online coupled chemical transport model simulations were analyzed for three summer months of 2015 in Poland. One of them was run with default emission inventory, the other two with NOx and VOC emissions reduced by 30%, respectively. Obtained ozone concentrations were evaluated with data from air quality measurement stations and ozone sensitivity to precursor emissions was estimated by ozone concentration differences between simulations and with the use of indicator ratios. They were calculated based on modeled mixing ratios of ozone, total reactive nitrogen and its components, nitric acid and hydrogen peroxide. The results show that the model overestimates ozone concentrations with the largest errors in the morning and evening, which is primarily related to the way vertical mixing is resolved by the model. Better model performance for ozone is achieved in rural than urban environment, as PBL and mixing mechanisms play more significant role in urban areas. Modeled ozone shows mixed sensitivity to precursor concentrations, similarly to other European regions, but indicator ratios have different values than are found in literature, particularly H2O2/HNO3 is larger than in southern Europe. However, indicator ratios often differ between locations and transition values need to be established individually for a given region.