In this Perspective article, we explore the definition and use of clinical tolerability metrics associated with nonspecific eddy current heating in magnetic field hyperthermia (MFH). We revisit the origins of the “Brezovich criterion,” Hof ≤ 485 MA m−1s−1, as it is applied to axial time-varying magnetic fields H (t) = Ho sin(2πft) and the human torso. We then consider alternative metrics, including the “maximal specific absorption rate” (SARmax) of eddy-current-induced power absorbed per unit mass of tissue. With reference to previously published clinical data and the results of two volunteer studies in our laboratory, we show that the SARmax metric is both suitable and reliable. We also show how it may be extracted from in silico finite element models to cope with confounding effects such as anatomical hot spots and non-axial-field geometries. We note a parallel with a standardized metric, the “local SAR” used in magnetic resonance imaging (MRI). We suggest that the limits established in clinical MRI (that the local SAR, averaged over 10 g of tissue and 6 min of treatment, should not exceed 20 mW g−1 in the torso or head, and 40 mW g−1 in the limbs) might be regarded as a good starting point for the design of MFH interventions. We conclude with the recommendation that the SARmax metric is adopted for future use in the development of clinically safe and tolerable MFH equipment.