Hypervelocity impacts on terrestrial bodies have the potential to rapidly heat and redistribute silicate target material to form impact melt flows. On the Moon, a subset of impact melt deposits exited the craters as laterally moving flows that moved away from the crater rim under the influence of gravity. These impact melt flows exhibit similarities to lava flows, but have the potential to be superheated by the impact cratering process. To estimate the initial temperature (T0) and thickness (H0) of these flows we combine new remote sensing analyses, experimentally-derived rheological relationships for three lunar analog materials, and numerical forward models of impact melt flow emplacement to identify combinations of parameters that yield flows matching the observed length of impact melt deposits on the Moon. We focus on the impact melt flows of four craters, which span a range of target materials and crater diameters (D): Giordano Bruno (D = 22.1 km), Necho (D = 36.9 km), Tycho (D = 85.3 km), and Copernicus (D = 96.1 km). We modeled three melt compositions-anorthosite, norite, and basalt-emplaced under cold ambient temperatures (Tambient = -180 ° C) on the Moon, to place a minimum bound on cooling-limited flow emplacement. Results show that observed flow lengths can be achieved by subliquidus melts emplaced within a laminar flow regime. For high-density (0% vesicularity) melts, best-fit inversions for T0 and H0 have mean values of 1253.3 ± 149.8 ° C and 21.6 ± 8.7 m, respectively; for low-density (64% vesicularity) melts, mean T0 and H0 values are 1343.3 ± 87.8 ° C and 27.9 ± 5.7 m-with uncertainties reported at 1 standard deviation. Model results also show that under cold ambient temperatures, impact melt flows with 0 to 64% vesicularity would be expected to have mean emplacement durations of 76, 78, 174, and 186 h for Giordano Bruno, Necho, Tycho, and Copernicus, respectively. Therefore, initial impact melt temperatures do not have to be in the super-liquidus range to form gravity-driven flow deposits matching observed lengths, and emplacement times are long enough for the majority of self-secondary fragments to land before the impact melt flows are fully emplaced. This suggests that some impact melt flow may be lava-like in their emplacement dynamics and that differences in crater populations on impact melt flows and adjacent ejecta may differ significantly, with impact melts providing a more robust estimate of the emplacement age of the young craters due to the presence of fewer secondaries. Impact melts could still have super-liquidus temperatures at the time of their initial formation, but turbulence would facilitate the rapid entrainment of external clasts, quickly lowering melt temperatures to generate flows that are similar to the products of effusive volcanic eruptions.