Sukhodolov et al. (2021) describes the configuration and some evaluation of the SOCOLv4 composition-climate model. It is generally very well written and presented, and I recommend its publication in GMD following some minor revisions. I have no major issues or comments to make on the manuscript.

The minor comments that I have are:

1. Page 3, lines 96-98: This is very minor - you introduce the component models and their acronyms here, but the references aren't given until Section 2. Objectively, this is fine as everything is referenced, but as a new reader I noticed that no references were given.
2. Figure 1: The colours used for the blue and turquoise boxes are quite similar, could you have a greater colour contrast?
3. Figure 1: Tropospheric aerosols are listed as a boundary condition for both ECHAM6 & MEZON, and I was wondering how these were different from the aerosols calculated from AER? Are these BCs purely the SADs for N2O5 hydrolysis? 
4. On page 5, lines 133-134 you state "Hereafter we refer to MPI-ESM1.2 as MPI-ESM", but then on page 11, line 296 you say "SOCOLv4 is based on the latest version of MPI-ESM (v1.2.01p6)". I'd suggest removing the version number here and putting it in the earlier section.
5. Figure 3b: given SOCOLv4 uses an interactive ocean, is a time-series comparison of Nino3.4 index the best method of comparing these quantities? Perhaps histograms similar to Figure 1 from Nowack et al. 2017 (https://doi.org/10.1002/2016GL072418) would be a better way to present these data?
6. Page 14, line 381: "The left panel of Fig. 4" - I would suggest ensuring that all figures have each sub-figure labelled (e.g. a, b, c etc.) and specific sub-plots or a range of sub-plots are referred to using these, e.g "Figures 4a - 4c..." etc. Some figures don't label their sub-plots (7, 8, 9, 12, A2), whereas other figures do (3, 4, 5, 6, 10, 11, 13, 14, 15, 16, 17, A1, A3). I would advise labelling all for consistency and clarity, and avoid using references similar to "the first plot in the left column" etc. 
7. Figures 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, A1, A2, & A3 do not have a colour-bar despite having coloured contours. While the black contours are labelled, I suggest also including the colourbar on these plots. 
8. Page 24, line 523 - what is being overestimated at the entry point? Is it the water vapour concentration itself, or something else?
9. Page 27, line 586: "CLONO₂" rather than ClONO₂
10. In Figure 13 I would be intested in seeing difference plots of total ozone, especially due to the stated reduction in TOC in SOCOLv4 coming from the corrected photolysis rates (page 31, lines 675-676).
11. The image quality of Figure 14 seems lower than the other figures, is it a lower quality raster image?
12. This is just a comment, but I found the differences in TOC coming from the aerosol climatology used for CCMI used and seen in Figure 15 very interesting.
13. Figure A2 caption: "ClNO₂" rather than ClONO₂.
14. Page 43, line 879: there is a gap in the DOI for the SOCOLv4 source code link
15. I would advise (although I'm sure you will do this in any case) to double-check references to pre-prints and discussion papers. For instance the Keeble et al. (2020) reference is now published, which occured after the discussion started for this paper.