A global three-dimensional model for studies of the tropospheric chemistry of HOx, ozone, and their precursors is thoroughly evaluated with available observations of 15 species from surface stations and/or aircraft campaigns. The effect of grid-resolution was studied by comparing the measurements to model runs at a high (~1.9¿ ¿ 1.9¿) and a reduced resolution. This study is a follow-up to a previous paper where the model was described and results for ozone were discussed. A statistical analysis of the complete comparison of each species is also presented in order to serve as a quantitative baseline for future evaluations. The seasonality and magnitude of the CO abundance is well simulated except for an overestimate of up to 20 nmol/mol in southern low latitudes. The tropospheric mean methane lifetime is calculated to be 8.7 and 9.1 years for the high- and low-resolution run, respectively, in agreement with recent estimates. Industrial emissions of alkanes from North America appear to be too low in the current data set. Indications for a biogenic source of propane were found by comparing with measurements over the Amazon rain forest. Alkene emissions from oceans are underestimated in the model indicating that they are higher than previous studies suggested. Nitrogen species are mostly reproduced within a factor of 2 or better, but a general bias to underestimate HNO3 and to overpredict peroxyacetyl nitrate (PAN), especially in remote regions could also be found. NO was well simulated in the lower and mid troposphere, but significantly understimated in the upper troposphere. A sensitivity run shows that studies are needed to better constrain the rate constants for the thermal equilibrium of PAN. Acetone is often underestimated indicating that additional wide-spread sources, likely oceanic, are present. The comparison with other oxygenated species (acetaldehyde, methanol, formic and acetic acid) indicates severe gaps in our understanding of the present budgets of these species. In particular the underestimate of acetaldehyde implies large missing sources (order 100 Tg/yr) of this compound.