Thank you submitting your revised manuscript which addressed most of the concerns of the referees. Following a further review from one of the referees, they have requested a few technical changes as follows:

Page 1, Line 13: This last sentence of the abstract sounds a bit strange. At least the “in” after "effective" should be deleted but I would suggest to reformulate the entire sentence. Do you mean that the drift is due to a stronger deep water formation around Antarctica in LIM3?

Page 4, Line 10: ETOPO1 (Amante et al., 2009)

Page 8, Line 18: “modeli”

Page 9, line 18: I think, the ice thickness should be the ratio between ice volume and ice area and not ice extent (otherwise ice free areas are also counted for ice concentrations > 15%).

Page 9, lines 27-31: If this is explaining the difference in trends in GIOMAS and LIM2/ LIM3 in the SH, maybe you should call it “GIOMAS” instead of “PIOMAS”.

Page 11, line 1: … faster than LIM2 ice motion.

Page 11, Sea ice salinity: It seems realistic that LIM3 simulates lower ice salinities in summer than in winter. Are there any observations of ice salinity available? If not, it might be worth to state this in the text.

Page 13, line 1: delete one “PHC3”

Page 13, line 26 …surface is saltier…

Page 16, Mixed Layer Depth: Mixed layer in NEMO might be deeper than in climatologies in many regions but are they really deeper in the deep water formation areas in Labrador Sea and Greenland Sea? It is hard/ impossible to see in the figure 9 because the figures are small but 9a) shows that NEMO-LIM3 has a mean March MLD of below 500m. If I look at the ARGO-climatology (Holte et al. 2016), I find March-mean values of up to 1000m in the Labrador Sea. To my understanding, the weak deep water formation, particularly in the Labrador Sea, in NEMO is the main reason for the weak AMOC in NEMO.

Editor comments:

The title is improved from the last version, but I do not like the word “some”, and also the order of the things you are comparing. The first thing you look at is sea ice, and it is the bulk of the paper, so this should come first in the list. Please revise.

I do not think the change to the colours in Fig.3 will improve the situation in relation to the issue of color-blindness, I suggest using different colours away from red and green, rather than just changing the tone as you have done.

The paper has a lot of acronymns, particularly on page 6, many of which are undefined, while some are common and may be assumed, I think in particular OPA on page 3, and PHC3 on page 6 should be spelt out.

page 16: line 12, should be “LIM3 becomes, on average, saltier”