Central nervous system (CNS) tumors are thought to be poorly immunogenic. However, whether defective anti-tumor immunity is a consequence of a relative failure of T cell priming versus a deficient effector phase of the anti-tumor immune response is not clear. We utilized a well-defined model system of B16 melanoma expressing the model antigen SIY-GFP to evaluate tumor antigen cross-priming and tumor rejection from the CNS versus subcutaneous compartments. We observed that B16-SIY cells implanted in the CNS were capable of inducing T cell priming as measured by IFN-gamma ELISPOT in the spleen. Cross-priming occurred in the absence of detectable systemic dissemination of the tumor. Despite the induction of a T cell response, CNS tumors grew progressively and were fatal, whereas the same tumor implanted in the flank was rejected. To study the effector phase of the immune response in more detail, in vitro primed 2C/RAG2-/- TCR transgenic CD8+ cells, which recognize the SIY peptide, were adoptively transferred. In addition, the CNS microenvironment was modulated by intracranial delivery of IL-2. While mice that received primed 2C cells alone showed an increase in survival, co-administration of intracranial IL-2 led to a marked prolongation of survival, with 20% of mice surviving at least 120 days. Our results demonstrate that CD8+ T cell cross-priming does indeed occur in response to a CNS tumor, but that manipulation of the brain tumor microenvironment may be necessary to support the effector phase of the anti-tumor immune response.