There are currently several recommended drug regimens for uncomplicated falciparum malaria in Africa. Each has different properties that determine its impact on disease burden. Two major antimalarial policy options are artemether-lumefantrine (AL) and dihydroartemisinin-piperaquine (DHA-PQP). Clinical trial data show that DHA-PQP provides longer protection against reinfection, while AL is better at reducing patient infectiousness. Here we incorporate pharmacokinetic-pharmacodynamic factors, transmission-reducing effects and cost into a mathematical model and simulate malaria transmission and treatment in Africa, using geographically explicit data on transmission intensity and seasonality, population density, treatment access and outpatient costs. DHA-PQP has a modestly higher estimated impact than AL in 64% of the population at risk. Given current higher cost estimates for DHA-PQP, there is a slightly greater cost per case averted, except in areas with high, seasonally varying transmission where the impact is particularly large. We find that a locally optimized treatment policy can be highly cost effective for reducing clinical malaria burden.