The amount of MgATP hydrolyzed per pair of electrons transferred (ATP/2e) during nitrogenase catalysis (1.0 atm N2, 30 °C) using titanium(III) citrate (Ti(III))as reductant was measured and compared to the same reaction using dithionite (DT). ATP/2e values near 210 for Ti(III) and 5.0 for DT indicate that nitrogenase has a much lower ATP requirement using Ti(III) as reductant. Using reduced Azotobacter vinelandii flavoprotein (AvFlpH2), a possible in vivo nitrogenase reductant, ATP/2e value near 2.0 were also observed. When the reaction was conducted using Ti(III) under N2, 5% CO in N2, Ar, 5% CO in Ar, or acetylene, ATP/2e values near 2.0 were also observed. With Ti(III) as reductant, ATP/2e values near 2.0 were measured as a function of temperature, Fe:MoFe protein ratio, and MoFe:Fe protein ratio, in contrast to measured values of 4.0-25 when DT is used under the same conditions. Both Ti(III) and AvFlpH2 are capable of forming the [Fe4S4]0 cluster state of the Fe protein whereas DT is not, suggesting that ATP/2e values near 2.0 arise from operation of the [Fe4S412+/[Fe4S4]0 redox couple with hydrolysis of only 2 ATPs per pair of electrons transferred. Additional experiments showed that ATP/2e values near 2.0 correlated with slower rates of product formation and that faster rates of product formation produced ATP/2e values near 5.0. ATP/2e values of 5.0 are consistent with the operation of the [Fe4S4]2+/[Fe4S4]1+ redox couple while ATP/2e values of 2.0 could arise from operation of the [Fe4S4]2+/[Fe4S4]0 redox couple. These results suggest that two distinct Fe protein redox couples may be functional during nitrogenase catalysis and that the efficiency of ATP utilization depends on which of these redox couples is dominant.