The intracellular parasite Toxoplasma gondii infects a wide variety of vertebrate species globally. Infection in most hosts causes a lifelong chronic infection and generates immunological memory responses that protect the host against new infections. In regions where the organism is endemic, multiple exposures to T. gondii likely occur with great frequency, yet little is known about the interaction between a chronically infected host and the parasite strains from these areas. A widely used model to explore secondary infection entails challenge of chronically infected or vaccinated mice with the highly virulent type I RH strain. Here, we show that although vaccinated or chronically infected C57BL/6 mice are protected against the type I RH strain, they are not protected against challenge with most strains prevalent in South America or another type I strain, GT1. Genetic and genomic analyses implicated the parasite-secreted rhoptry effectors ROP5 and ROP18, which antagonize the host's gamma interferon-induced immunity-regulated GTPases (IRGs), as primary requirements for virulence during secondary infection. ROP5 and ROP18 promoted parasite superinfection in the brains of challenged survivors. We hypothesize that superinfection may be an important mechanism to generate T. gondii strain diversity, simply because two parasite strains would be present in a single meal consumed by the feline definitive host. Superinfection may drive the genetic diversity of Toxoplasma strains in South America, where most isolates are IRG resistant, compared to North America, where most strains are IRG susceptible and are derived from a few clonal lineages. In summary, ROP5 and ROP18 promote Toxoplasma virulence during reinfection. Toxoplasma gondii is a widespread parasite of warm-blooded animals and currently infects one-third of the human population. A long-standing assumption in the field is that prior exposure to this parasite protects the host from subsequent reexposure, due to the generation of protective immunological memory. However, this assumption is based on clinical data and mouse models that analyze infections with strains common to Europe infections with strains common to Europe and North America. In contrast, we found that the majority of strains sampled from around the world, in particular those from South America, were able to kill or reinfect the brains of hosts previously exposed to T. gondii. The T. gondii virulence factors ROP5 and ROP18, which inhibit key host effectors that mediate parasite killing, were required for these phenotypes. We speculate that these results underpin clinical observations that pregnant women previously exposed to Toxoplasma can develop congenital infection upon reexposure to South American strains.