This paper studies the gravitational wave power spectra formed by instantaneously initiated magnetohydrodynamic turbulence. Informed by a combination of new simulations and previously established properties of the shear stress tensor, the authors develop an analytical model of the gravitational wave power spectrum and demonstrate agreement with their simulation results. Their focus is the low-wavenumber behaviour of the power spectrum, motivated by the possibility of an inverse cascade transferring power to longer length scales. The paper is interesting. The work addresses important issues, representing real progress in this area of research. I will be able to recommend publication after the following issues have been addressed: (1) The final four paragraphs of the Introduction ("To address the problem..." onward) mix elements of a literature review and the motivation for the present work in a way which is not quite successful at either. I suggest the authors try to make clearer how the existing literature has informed the present work, and to make clearer what is being done here for the first time. (2) It is also not clear from the Introduction what the relationship is with the simulations of Ref. 23. The authors state that "we also provide the evolution of the stress spectrum for these simulations [presumably those in Ref. 23]". This would seem to imply that the simulations in the present work are the same or have the same parameters as those in Ref. 23, or that the results of those simulations are reanalyzed in the current work, but as far as I can tell that is not the case. Answer- A table in appendix with our parameter and those with Alberto Roper Pol (3) Many of the plots lack clear legends, which is an impediment to understanding the results of the paper. It is not at all clear what the different colors mean in Figures 1 and 2 - presumably different times. The colors for Figures 3 and 4 are given in the caption text, but that may not be sufficient for color-blind readers. Even with color vision and a color printer, it is hard to distinguish the curves in Figure 5; in this case it would be good to explain in the caption which curves exactly coincide with which other curves. The caption in Figure 8, left, is not optimally placed while Figure 8, right, appears to lack a k_p=1 curve despite it being present in the legend. Figure 12 has arrows identifying each curve, but the caption again relies on colors that are difficult to interpret. Answer- In figure 5, write t on the curves with arrow. Oragne-> Blue dashed (4) The model of Eq. 5, while undoubtedly leading to good agreement (as seen in Figure 12), does not seem to be supported by the behaviour of \tilde{T} for the helical case (seen in Figure 3, top left) which does not seem to have a clear power-law. Could the authors perhaps comment on how the model nevertheless seems to work? They may also wish to expand the discussion in the third paragraph of the left column on page 3, left, where these results are introduced. That would allow a better motivation of the model in Eq. 5. They could also extend their results in Figures 3 and 4 to the k=3 mode shown in Fig. 11 which appears to have a clearer increase for both cases. Response- Investigate top left panel of Figure (3) and think about the comparison with the plot in figure (11) left panel. (5) The time at which the results shown in Figure 12 are plotted does not appear to be stated everywhere. Is agreement consistently good at all times? Response- Make a plot with a particular wavenumber for GW spectrum at different time I have some further minor points: - In the Introduction, the authors write about "the various phase transitions in the early Universe" but we have not established that any such phase transitions took place (as far as we know, the QCD and electroweak transitions were crossovers). - The run labels 'hel' and 'nonhel' are sometimes hard to pick out of the text, perhaps they could be typeset in a different typeface. Furthermore, in some cases they are instead referred to as the "helical and nonhelical runs" (e.g. in the text two paragraphs below Table I) or as "... with magnetic helicity" and "... for the nonhelical case". These should be consistently labeled throughout the paper, and I would suggest erring on the side of verbosity (thus: "... for the nonhelical (HEL) case"). Response- HEL and NHEL - Table II has column headings 'p' but it is not clear that what is actually in the table is the power p but rather the overall time dependence of the shear stress.