Although the study of non-human primates has resulted in important advances for understanding HIV-specific immunity, a clear correlate of immune control over simian immunodeficiency virus (SIV) replication has not been found to date. In this study, CD8⁺ T-cell cytotoxic capacity was examined to determine whether this function is a correlate of immune control in the rhesus macaque (RM) SIV infection model as has been suggested in chronic HIV infection. SIVmac251-infected human reverse transcriptase (hTERT)-transduced CD4⁺ T-cell clone targets were co-incubated with autologous macaque effector cells to measure infected CD4⁺ T-cell elimination (ICE). Twenty-three SIV-infected rhesus macaques with widely varying plasma viral RNA levels were evaluated in a blinded fashion. Nineteen of 23 subjects (83%) were correctly classified as long-term nonprogressor/elite controller (LTNP/EC), slow progressor, progressor or SIV-negative rhesus macaques based on measurements of ICE (weighted Kappa 0.75). LTNP/EC had higher median ICE than progressors (67.3% [22.0–91.7%] vs. 23.7% [0.0–58.0%], p = 0.002). In addition, significant correlations between ICE and viral load (r = −0.57, p = 0.01), and between granzyme B delivery and ICE (r = 0.89, p<0.001) were observed. Furthermore, the CD8⁺ T cells of LTNP/EC exhibited higher per-cell cytotoxic capacity than those of progressors (p = 0.004). These findings support that greater lytic granule loading of virus-specific CD8⁺ T cells and efficient delivery of active granzyme B to SIV-infected targets are associated with superior control of SIV infection in rhesus macaques, consistent with observations of HIV infection in humans. Therefore, such measurements appear to represent a correlate of control of viral replication in chronic SIV infection and their role as predictors of immunologic control in the vaccine setting should be evaluated.