Ellipticine is a DNA-damaging agent acting as a prodrug whose pharmacological efficiencies and genotoxic side effects are dictated by activation with cytochrome P450 (CYP). Over the last decade we have gained extensive experience in using pure enzymes and various animal models that helped to identify CYPs metabolizing ellipticine. In this review we focus on comparison between the in vitro and in vivo studies and show a necessity of both approaches to obtain valid information on CYP enzymes contributing to ellipticine metabolism. Discrepancies were found between the CYP enzymes activating ellipticine to 13-hydroxy- and 12-hydroxyellipticine generating covalent DNA adducts and those detoxifying this drug to 9-hydroxy- and 7-hydroellipticine in vitro and in vivo. In vivo, formation of ellipticine-DNA adducts is dependent not only on expression levels of CYP3A, catalyzing ellipticine activation in vitro, but also on those of CYP1A that oxidize ellipticine in vitro mainly to the detoxification products. The finding showing that cytochrome b5 alters the ratio of ellipticine metabolites generated by CYP1A1/2 and 3A4 explained this paradox. Whereas the detoxification of ellipticine by CYP1A and 3A is either decreased or not changed by cytochrome b5, activation leading to ellipticine-DNA adducts increased considerably. We show that (I) the pharmacological effects of ellipticine mediated by covalent ellipticine-derived DNA adducts are dictated by expression levels of CYP1A, 3A and cytochrome b5, and its own potency to induce these enzymes in tumor tissues, (II) animal models, where levels of CYPs are either knocked out or induced are appropriate to identify CYPs metabolizing ellipticine in vivo, and (III) extrapolation from in vitro data to the situation in vivo is not always possible, confirming the need for these animal models.