n northern US and Europe thermal cracking is a dominant distress for asphalt pavements. For this reason, a robust and reliable characterization method of asphalt binder at low temperature is fundamental for correctly selecting the right material and for design of asphalt mixtures which can provide satisfactory long-term performance. In the US, this issue is addressed based on the Performance Grading (PG) system, and on low temperature creep tests performed on asphalt binder with the Bending Beam Rheometer (BBR). However, this test relies on a cooling bath consisting of ethanol, which does not entirely represent the environmental condition of the pavement in the field and which may ultimately provide misleading results. In this paper, the possibility of using air as an alternative cooling medium for testing asphalt binder in the BBR is evaluated. For this purpose, five asphalt binders were characterized with the BBR; creep stiffness, m-value, PG, thermal stress and critical cracking temperature were computed both for ethanol and air. In addition, the rheological Huet model was fitted to the experimental measurements to further investigate the effect of the cooling medium. It was found that air measurements result in stiffer materials, with higher low PG, higher thermal stress and critical cracking temperature. The parameters of the Huet model confirm such a stiffening effect when air is used. Based on the material response observed, it is recommended to use air in place of ethanol in the BBR, as the latter provides an overestimation of the material properties.