Abstract. We validate OMI ozone profiles between 0.22–215 hPa and stratospheric ozone columns down to 215 hPa (SOC215) against v2.2 MLS data from 2006. The validation demonstrates convincingly that SOC can be derived accurately from OMI data alone, with errors comparable to or smaller than those from current MLS retrievals, and it demonstrates implicitly that tropospheric ozone column can be retrieved accurately from OMI or similar nadir-viewing ultraviolet measurements alone. The global mean biases are within 2.5% above 100 hPa and 5–10% below 100 hPa; the standard deviations of the differences (1σ) are 3.5–5% between 1–50 hPa, 6–9% above 1 hPa and 8–15% below 50 hPa. OMI shows some latitude and solar zenith angle dependent biases, but the mean biases are mostly within 5% and the standard deviations are mostly within 2–5% except for low altitudes and high latitudes. The excellent agreement with MLS data shows that OMI retrievals can be used to augment the validation of MLS and other stratospheric ozone measurements made with even higher vertical resolution than that for OMI. OMI SOC215 shows a small bias of −0.6% with a standard deviation of 2.8%. When compared as a function of latitude and solar zenith angle, the mean biases are within 2% and the standard deviations range from 2.1 to 3.4%. Assuming 2% precision for MLS SOC215, we deduce that the upper limits of random-noise and smoothing errors for OMI SOC215 range from 0.6% in the southern tropics to 2.8% at northern middle latitudes.