Three COVID-19 vaccines have received FDA-authorization and are in use in the United States, but there is limited head-to-head data on the durability of the immune response elicited by these vaccines. Using a quantitative assay we studied binding IgG antibodies elicited by BNT162b2, mRNA-1273 or Ad26.COV2.S in an employee cohort over a span out to 10 months. Age and sex were explored as response modifiers. Of 234 subjects in the vaccine cohort, 114 received BNT162b2, 114 received mRNA-1273 and six received Ad26.COV2.S. IgG levels measured between seven to 20 days after the second vaccination were similar in recipients of BNT162b2 and mRNA-127 and were ~50-fold higher than in recipients of Ad26.COV2.S. However, by day 21 and at later time points IgG levels elicited by BNT162b2 were lower than mRNA-1273. Accordingly, the IgG decay curve was steeper for BNT162b2 than mRNA-1273. Age was a significant modifier of IgG levels in recipients of BNT162b2, but not mRNA-1273. After six months, IgG levels elicited by BNT162b2, but not mRNA-1273, were lower than IgG levels in patients who had been hospitalized with COVID-19 six months earlier. Similar findings were observed when comparing vaccine-elicited antibodies with steady-state IgG targeting seasonal human coronaviruses. Differential IgG decay could contribute to differences observed in clinical protection over time between BNT162b2 and mRNA-1273.