Measurements of OH, HO2, RO2i (alkene and aromatic-related RO2) and total RO2 radicals taken during the ClearfLo campaign in central London in the summer of 2012 are presented. A photostationary steady-state calculation of OH which considered measured OH reactivity as the OH sink term and the measured OH sources (of which HO2+ NO reaction and HONO photolysis dominated) compared well with the observed levels of OH. Comparison with calculations from a detailed box model utilising the Master Chemical Mechanism v3.2, however, highlighted a substantial discrepancy between radical observations under lower NOx conditions ([NO] < 1 ppbv), typically experienced during the afternoon hours, and indicated that the model was missing a significant peroxy radical sink; the model overpredicted HO2 by up to a factor of 10 at these times. Known radical termination steps, such as HO2 uptake on aerosols, were not sufficient to reconcile the model-measurement discrepancies alone, suggesting other missing termination processes. This missing sink was most evident when the air reaching the site had previously passed over central London to the east and when elevated temperatures were experienced and, hence, contained higher concentrations of VOCs. Uncertainties in the degradation mechanism at low NOx of complex biogenic and diesel related VOC species, which were particularly elevated and dominated OH reactivity under these easterly flows, may account for some of the model-measurement disagreement. Under higher [NO] (> 3 ppbv) the box model increasingly underpredicted total [RO2]. The modelled and observed HO2 were in agreement, however, under elevated NO concentrations ranging from 7 to 15 ppbv. The model uncertainty under low NO conditions leads to more ozone production predicted using modelled peroxy radical concentrations ( ≈ 3 ppbv h-1) versus ozone production from peroxy radicals measured ( ≈ 1 ppbv h-1). Conversely, ozone production derived from the predicted peroxy radicals is up to an order of magnitude lower than from the observed peroxy radicals as [NO] increases beyond 7 ppbv due to the model underprediction of RO2 under these conditions.