Artemisinin-based drugs are the front-line weapon in the treatment of human malaria cases, but there is concern that recent reports of slow clearing infections may signal developing resistance to treatment. In the absence of molecular markers for resistance, current efforts to monitor drug efficacy are based on the rate at which parasites are cleared from infections. However, some knowledge of the standing variation in parasite susceptibility is needed to identify a meaningful increase in infection half-life. Here, we show that five previously unexposed genotypes of the rodent malaria parasite Plasmodium chabaudi differ substantially in their in vivo response to treatment. Slower clearance rates were not linked to parasite virulence or growth rate, going against the suggestion that drug treatment will drive the evolution of virulence in this system. The level of variation observed here in a relatively small number of genotypes suggests existing ‘resistant’ parasites could be present in the population and therefore, increased parasite clearance rates could represent selection on pre-existing variation rather than de novo resistance events. This has implications for resistance monitoring as susceptibility may depend on evolved traits unrelated to drug exposure.