Pyrrolizidine alkaloids (PAs) are toxic plant constituents occurring often in their N-oxide form. This raises the question on the relative potency (REP) values of PA-N-oxides compared to the corresponding parent PAs. The present study aims to quantify the in vivo REP value of riddelliine N-oxide compared to riddelliine using physiologically based kinetic (PBK) modelling, taking into account that the toxicity of riddelliine N-oxide depends on its conversion to riddelliine by intestinal microbiota and in the liver. The models predicted a lower C-max and higher T-max for the blood concentration of riddelliine upon oral administration of riddelliine N-oxide compared to the C-max and T-max predicted for an equimolar oral dose of riddelliine. Comparison of the area under the riddelliine concentration-time curve (AUC(RID)) obtained upon dosing either the N-oxide or riddelliine itself revealed a ratio of 0.67, which reflects the in vivo REP for riddelliine N-oxide compared to riddelliine, and appeared to closely match the REP value derived from available in vivo data. The models also predicted that the REP value will decrease with increasing dose level, because of saturation of riddelliine N-oxide reduction by the intestinal microbiota and of riddelliine clearance by the liver. It is concluded that PBK modeling provides a way to define in vivo REP values of PA-N-oxides as compared to their parent PAs, without a need for animal experiments.