Measurements of flame length are reported that help to quantify the overall fuel-air mixing process that occurs within a jet-like flame in the supersonic regime. A nonpremixed, turbulent, hydrogen-air jet flame is stabilized along the axis of a Mach 2.2 coflowing air stream. Supersonic flame lengths are compared to measured lengths of subsonic hydrogen-air flames with coflow that were stabilized on the axis of a subsonic wind tunnel for a range of fuel/air velocity and density ratios. The supersonic dames are found to be significantly shorter than (i.e., typically half as long as) corresponding subsonic flames, providing that both the velocity and density ratios are matched for the two cases. This difference implies that for axisymmetric jet geometries with combustion, mixing rates are larger in the supersonic case. Compressibility effects art not believed to be significant since typical convective Mach numbers are less than 0.45. One possible reason why the supersonic flames are relatively short is that the radial velocity, which controls entrainment, can be altered by compression/expansion waves that are inherent to the jet geometry. Both subsonic and supersonic flame lengths increase as the normalized fuel mass flux increases, which is predicted to occur by a scaling analysis. The density of the supersonic airstream was decreased in order to simulate an increase in altitude, and the supersonic flames became longer, as predicted by the analysis. Increasing the air stagnation temperature to 600K shortens the supersonic flames, which is not explained by the mixing-limited scaling analysis and may be due to finite-rate chemistry or partially premixed combustion, both of which are temperature dependent processes.