This manuscript sets out to describe the new UK Met Office LFRic modelling framework as applied to the general problem of simulating atmospheric circulations that may be well removed from that of present day Earth itself. The detailed formulation is mostly presented fairly thoroughly with plenty of references and the manuscript goes on to present a set of simulations of some well known test cases for planetary atmosphere modelling for comparison with results of other GCMs used recently for exoplanet studies. In general the results seem to be encouraging in demonstrating that LFRic-Atmosphere produces results that are largely consistent with predecessor GCMs (including the current Met Office UM) for most test cases and satisfies some important tests of conservation of key integral quantities such as mass and angular momentum. The results therefore confirm that LFRic-Atmosphere has the potential to be a valuable new tool for planetary and exoplanetary studies, offering the possibility of interfacing it to some quite sophisticated parameterisation schemes for physical and chemical processes. The addition of the Trappist-1 test cases are particularly interesting and would merit further more detailed analysis - though perhaps for another publication that focuses more on scientific results than on the modelling methods.

The manuscript itself seems to be generally well written and organised. It provides much useful detail and background on the model code itself, which has a number of unusual and innovative features. The test cases seem generally well chosen and make for useful and convincing comparisons with the results of similar tests with other GCM codes. The manuscript could be accepted more or less as it is, though I have listed below a few points that the authors can respond to in a revised version. 

Major point:

One of the more significant points concerns the choice of the cubed sphere grid. An earlier intercomparison of exoplanetary GCM codes by Polichtchouk et al. (2014) indicated that the cubed sphere version of MITgcm performed least well in some test cases than other discretisation methods, citing issues with conservation properties and other artefacts related to the grid. It may be helpful to include a brief discussion of why LFRic-Atmosphere does not seem to display these kinds of issue compared with MITgcm.

Other minor points:

Line 15 - the use of the word “precipice” here may not carry the meaning intended by the authors. Moving beyond a precipice has the sense of falling off a cliff, with the natural (somewhat catastrophic!) consequences! Perhaps “threshold” might be a more auspicious word choice here?

Line 102 - The neglect of latitudinal variations in geopotential ignores changes in g between equator and pole? This is significant at the 0.5% level for Earth (and is probably bigger on fast-rotating gas giants?). 

Line 109 - Perhaps a good place to discuss the choice of cubed sphere in comparison with Polichtchouk et al 2014?

Eqs (50), (6) and (11) - why split these into 2 lines? Seems unnecessary and leads to potentially confusing disparity in sizes of brackets.

Lines 279-80 - You could use a dimensionless measure of AM such as in Lewis et al. (2021. Characterizing Regimes of Atmospheric Circulation in Terms of Their Global Superrotation, J. Atmos Sci., 78, 1245-58 and references therein)? 

Line 347 - Perhaps helpful to emphasise that clouds and microphysics here refer only to water (exoplanets max have clouds of varying composition!). 

Lines 352-3 - Perhaps give references for details of GA7.0 and GA9.0 configurations?

Figure 8 and associated text - Zonal mean fields are not necessarily very illuminating for tidally-locked planets. It is perhaps beyond the scope of this paper, but a decomposition following Hammond & Lewis 2021 may be more enlightening? 

Line 497 - “While we cannot judge which THAI GCM is more correct due to the absence of observations” - which is the bane of almost all exoplanet circulation studies! But more generally it may be useful to include a statement emphasising what new advantages LFRic-Atmosphere offers to the planetary atmosphere modelling community compared with other codes. Some of this is covered in the Introduction, but may be worth emphasising in the conclusions. 

References - several references display the titles of articles entirely as upper case, which looks strange. 

General comments:

The article “Simulations of idealised 3D atmospheric flows on terrestrial planets using LFRic-Atmosphere” by Sergeev et al. presents seven terrestrial planet benchmarks from the LFRic model, an evolution of the UK Met Office’s Unified Model (UM). The paper describes briefly the changes to the model from previous versions of the UM, then applies the model to three temperature forcing (TF) benchmarks, and finally to the four THAI cases. The authors compare each simulation qualitatively and quantitatively to published results from other models and to past UM results, and analyze the causes of differences between the UM and LFRic results using incremental changes to the model configurations. It is well-written with clear, readable figures, and provides a necessary step toward further application of this new model to terrestrial planets. I have only minor comments and questions, largely related to clarity. 

Specific comments:

1. Equation 1a-d: is there an equation for water vapor transport that is formally part of GungHo? Because the THAI Hab 1 and Hab 2 cases presumably have moist dynamics, it is worth explaining a bit here how moisture is treated in the model. 
2. Equation 1c: I noticed that the diabatic heating term is not included here. I am familiar enough with this type of modeling to know that it is often understood that the heating term will be added when radiative transfer, boundary layer processes, etc., are included, but some readers may not know this. So I suggest that it is worth adding that term (like in Mayne et al. 2014) or adding a note in the following paragraph regarding diabatic heating. (This is explained later for the temperature forcing cases, but not the THAI cases).
3. L115: “a necessary condition for avoiding computational modes”: Could the authors explain what is meant by “computation modes”, and how this condition avoids them?
4. L117: “with the mesh treated as structured in the vertical (radial) direction.”: this phrase is unclear to me. Are the authors referring to how the data is organized in memory (as in the following sentence), or something else? Please clarify.
5. Fig 1: does the right figure show the total (scalar) wind speed, or the velocity in a particular direction (zonal, e.g.)? Please clarify.
6. Page 6, footnote 1: explain why the W0 and W1 modes are not used in the current work.
7. L171-173: I presume the temperature forcing cases all use “dry” dynamics (no water vapor transport, no moist convection, no clouds, etc.). Please state this for clarity.
8. Equation 2: should there be a minus sign in front of the second part of the equation? I.e., “- (T-Teq)/tau_rad”
9. L184: “after which we assume it has reached a statistically steady state”: it seems to me that the evolution of Total Kinetic Energy provides a sufficient indicator of “spin up”, so that you do not need to “assume” it has reached steady state. I suggest rephrasing this sentence to show that you are confident in the model reaching steady state by this time and don’t need to make an assumption.
10. Fig 2: the conservation of mass appears to be quite good, though this could be simply because the total mass is so much larger than the error that the error is invisible on this scale. You should enlarge the y-scale on this figure so that the dynamic range is visible or include a value in the text indicating how large the error in mass is (e.g., 1 part in 10^x). This will be a helpful point of reference for future LFRic users/developers and for developers of other GCMs. 
11. Equations 5 & 11: Please indicate whether the “log” is base 10 or the natural logarithm, as there are different conventions in literature that can make it ambiguous. 
12. L228-229: “the dominant jets are only 3 m s-1 slower in LFRic-Atmosphere than in the UM”: in fact, this slight reduction in wind speed brings the result closer to the results in Held & Suarez and Heng et al 2011. Whether that really constitutes an “improvement” over the UM is a bit subjective as there is naturally some spread in this result, but I think it is worth making note of it. 
13. L245: since LFRic is using an altitude grid, I am guessing that sigma_stra varies depending on the current pressure at z_stra. Is that correct? Please clarify.
14. L271: was the sponge layer unnecessary (and not used) in the previous TF cases (and in the UM TF cases)? Please clarify.
15. L279: “for display purposes AAM is multiplied by 365…”: you should also explain this in the caption of Fig 2. 
16. L295, “LFRic-Atmosphere predicts equatorial regions of counter-flow in the zonal wind (Fig. 5b), while it is purely divergent from the substellar point in Merlis and Schneider (2010, Fig. 4a).”: I believe that the Merlis & Schneider study included a hydrological cycle, which is neglected in the Heng et al. study and (I believe) in yours. Could this be the reason for the difference in zonal flow?
17. L328-329: “an aquaplanet with infinite water supply (slab ocean)”: the phrase seems to imply that “slab ocean” means “infinite water supply”, but these are distinct concepts. A slab ocean simply means that the ocean is represented only by diffusive heat transport, rather than resolved dynamics. I don’t think that you intended to equate these concepts, but I suggest rephrasing to be more clear.
18. L328-329, again: “an aquaplanet with infinite water supply (slab ocean)”: does the infinite water supply become relevant in the Hab 1 or Hab 2 cases? I.e., what is the mass of water vapor that ends up in the atmosphere, compared to, say, an Earth ocean. I suppose the point of this is to indicate that you are not simulating “Dune” worlds with limited water inventories, a la Abe et al. 2011.
19. L332: “roughness length is set to…”: these are related to the boundary layer, correct? Please explain that, and provide a citation to a boundary layer description, if so. 
20. L334-335: Thank you for explaining how you determine steady state. I find statements like this are infrequent in GCM papers, but they are important for context and comparison of results. 
21. L342-343: “Note that the model top is lower than that used in the UM simulations…” Why did you use a lower model top in this work? Is it because of stability issues or some other constraint? Please explain. Also, could this be contributing to the differences in zonal flow between the UM and LFRic? 
22. Figure 6: Does the “net upward LW flux” refer to the upward surface to atmosphere LW flux (not including the downward atmosphere to surface beam), or to the net LW flux (e.g., F_up - F_down) with convention of positive upward, or something else? I suppose I am tripping over the combination of the words “net” and “upward”. Could you provide a mathematical definition in the text?
23. L412-414: Here, you state that the difference in circulation regime of Ben 1 between LFRic and UM is due to the dynamical core, but in the previous paragraph you state that re-running the UM with GA9.0 “leads to the regime change in the Ben 1 case”. I understood this to mean that it matches the LFRic result, which would indicate that the change to GA9.0 (rather than the dynamical core) is the cause of the difference between the older result and the newer one. Can you check this and clarify which change it is that matters?
24. L426-427, regarding the reduction in cloud ice: this seems like a significant difference between the model and is worth investigating. For example, could it be caused by the change to the dynamical core, numerical stability algorithms, or GA7.0 to GA9.0? The last change could be verified by comparing to the UM with GA9.0, which I believe you already have on hand. 
25. L442-444: Figure 7c shows a horizontal perspective of the winds, but Fig 10d in Sergeev et al. 2022 is the zonal-vertical perspective. So it isn’t easy to compare the overall pattern discussed here. Do you mean 9d in Sergeev 2022?
26. L449-451: it is a bit hard to understand fully without the zonal wind plot from the UM GA9.0 run, but based on the text description here, it sounds like that run agrees more with the earlier UM run in Sergeev et al 2022 than with the current LFRic result. Wouldn’t this indicate that the change in Hab 1’s zonal wind is due to GungHo, rather than the change to GA9.0? Please clarify. 
27. Figure 8: following on from the previous comment–because you compare so frequently to the UM GA9.0 run in the zonal winds, I think it is worth including those plots here, as a second row perhaps. This would make the comparison much easier to see. 
28. L474-475: not to suggest that the UM is “wrong” here, but I think it is worth stating that the LFRic results are closer to the results of the other three THAI GCMs. 

Technical corrections:

1. L46: I think a comma is missing in this sentence: “platforms making” -> “platforms, making” 
2. L303: “...places it on the edge of between two distinct…”: I think there is an extra word or two here, probably inserted during revision. Remove either “edge of” or “between”.
3. L462-463: “along the equator” appears twice in this sentence.