The use of anti-spike (S) serologic assays as surrogate measurements of SARS-CoV-2 vaccine induced immunity will be an important clinical and epidemiological tool. The characteristics of a commercially available anti-S antibody assay (Roche Elecsys anti-SARS-CoV-2 S) were evaluated in a cohort of vaccine recipients. Levels were correlated with pseudotype neutralizing antibodies (NAb) across SARS-CoV-2 variants. We recruited adults receiving a two-dose series of mRNA-1273 or BNT162b2 and collected serum at scheduled intervals up to 8 months post-first vaccination. Anti-S and NAb levels were measured, and correlation was evaluated by (i) vaccine type and (ii) SARS-CoV-2 variant (wild-type, Alpha, Beta, Gamma, and three constructs Day 146*, Day 152*, and RBM-2). Forty-six mRNA vaccine recipients were enrolled. mRNA-1273 vaccine recipients had higher peak anti-S and NAb levels compared with BNT162b2 ( < 0.001 for anti-S levels; < 0.05 for NAb levels). When anti-S and NAb levels were compared, there was good correlation (all values ≥ 0.85) in both BNT162b2 and mRNA-1273 vaccine recipients across all evaluated variants; however, these correlations were nonlinear in nature. Lower correlation was identified between anti-S and NAb for the Beta variant (= 0.88) compared with the wild-type (WT) strain (= 0.94). Finally, the degree of neutralizing activity at any given anti-S level was lower for each variant compared with that of the WT strain, ( < 0.001). Although the Roche anti-S assay correlates well with NAb levels, this association is affected by vaccine type and SARS-CoV-2 variant. These variables must be considered when interpreting anti-S levels. We evaluated anti-spike antibody concentrations in healthy mRNA vaccinated individuals and compared these concentrations to values obtained from pseudotype neutralization assays targeting SARS-CoV-2 variants of concern to determine how well anti-spike antibodies correlate with neutralizing titers, which have been used as a marker of immunity from COVID-19 infection. We found high peak anti-spike concentrations in these individuals, with significantly higher levels seen in mRNA-1273 vaccine recipients. When we compared anti-spike and pseudotype neuralization titers, we identified good correlation; however, this correlation was affected by both vaccine type and variant, illustrating the difficulty of applying a "one size fits all" approach to anti-spike result interpretation. Our results support CDC recommendations to discourage anti-spike antibody testing to assess for immunity after vaccination and cautions providers in their interpretations of these results as a surrogate of protection in COVID-vaccinated individuals.