The manuscript ‘GEOCLIM7, an Earth System Model for multi-million years evolution of the geochemical cycles and climate’ by Maffre et al. provides a complete overview and description of the components in the GEOCLIM earth system model. Several new components are outlined in this updated version that allow users to modify the set-up to the desired GCM input fields and/or paleogeography and climate states. Lastly, the authors demonstrate the applicability of GEOCLIM7 with a simulation of the Turonian forced with transient changes to the orbital configuration.

I am very happy to see a concise summary of the GEOCLIM model that, I am sure, will serve as an informative reference for those interested in interpreting as well as running GEOCLIM models. The extensive list of geochemical and physical equations and their parameterizations that form the basis of GEOCLIM provides transparency and clarification. I am also pleased to see an extensive explanation of the methods used for model calibration and tuning. The new components (e.g customizable ocean boxes, flexible GCM-derived forcing fields) are noteworthy improvements, particularly for the construction of paleo-configurations as demonstrated in Section 4.

I am very positive about this manuscript. It is well-written, clearly organized and I support publication after the comments and suggestions below are addressed:

1. In Sec. 2 (model description) and Sec. 3 (boundary conditions), the authors outline equations and interpolation methods used in the model. I would like to see a clearer indication of what components/equations are new to GEOCLIM7 and what components were previously published, and where. The sections that contain descriptions of pre-existing GEOCLIM equations lack referencing. I recommend adding citations to papers and/or GEOCLIM versions where the components are first described, tested, and applied. Such complete referencing will not only serve as a historical record of GEOCLIM development but also to track the steps that have been taken to validate the model and evaluate its performance.
2. The flexible ocean boxes are a novel addition in which ocean circulation is obtained from GCM output and converted into exchange fluxes between GEOCLIM boxes using a new tool. Can you show that the tool is indeed able to accurately reconstruct the large-scale ocean circulation, in particular the vertical component (W^V) that is indirectly calculated? Perhaps a figure that compares the ocean fluxes in the GCM with that of an older standard GEOCLIM configuration and that using the new version? Would also be useful to  reference Section 4 (where the ocean boxes specification is tested) in Section 3.2.
3. Another innovation is the seafloor sediment routing scheme (Sec. 2.3.2). A schematic representation of the routing scheme would be helpful to visualize the mechanics of transport and deposition fluxes, or if this methodology was developed in a different study, cite the appropriate paper. The manuscript is missing a test case scenario that demonstrates the implication of this innovation. How does this improve the model?
4. If I understand correctly, two tools are designed to (1) convert ocean water exchange fluxes from GCM to GEOCLIM (BC_generator.py), and to (2) generate a river routing scheme (basinmap_editor.py). Great that these are made publicly available along with the GEOCLIM download. The scripts are clear and well annotated, but it is not clear from the manuscript text how one would use these to re-grid their own GCM outputs. As highlighted by the authors, the main improvement of GEOCLIM7 is that boundary conditions from any GCM (not just FOAM) can be used so a step-by-step outline of how users would go about this seems indispensable. For instance, you could provide a readme file or short manual along with these scripts to improve user experience? Or indicate what scripts are needed to perform each step in Figure 3?
5. Is the lithological information (Section 3.4.2) a new addition to GEOCLIM7? If it is, this calls for a sensitivity test to check how sensitive results are to choice of lithology, especially important for paleo-configs in which lithology is often unknown. If this is not a new addition of GEOCLIM7, cite the appropriate manuscript in which it was developed and tested.
6. The Turonian experiment nicely demonstrates the applicability of the new model components. However, I’d like to see an explanation or discussion about the multilinear interpolation between climate fields of variable pCO2 and orbits (as mentioned in Section 3.4.3). In particular, because the relation between pCO2 and global climate response is not linear and only two end-member CO2 values are tested. Considering this, it would be useful to know if (and by how much) the pCO2 in the ‘All processes’ experiment changes, i.e. how do modelled variations fall in the range of 560-1120 uatm?
7. I’m surprised to see how insensitive the system is to orbital forcing in the ‘Cont. fluxes’ experiment. The authors explain the weak response by the low absolute P weathering flux but this, in my opinion, is an incomplete assessment. Is it possible that orbital changes to the terrestrial P flux are subdued because of the annual averaged surface temperature and runoff fields, unable to capture nonlinear seasonal behaviours of weathering (e.g equations in Sec. 2.4)? Can you address the absence of seasons in the model and how that may impact the orbital results, for weathering but also ocean circulation?

I have also listed additional minor line-by-line comments in the attached file that require attention. Overall, the new model components presented here are a step forward for GEOCLIM. I'm looking forward to seeing future research to the long term biogeochemical cycling in a variety of (paleo)climate settings facilitated by these new developments. 

This paper presents a technical description of the latest GEOCLIM model. It demonstrates the model calibration for present day and explores orbital-timescale changes during the Cretaceous.

The paper is informative and generally well explained. It will certainly be useful as an accompaniment to shorter form papers presenting specific tests of this model.

I found no major issues in the paper, but the figures for the orbital-timescale runs were difficult to interpret as they lacked y axis labels. I also thought it was a little disappointing that the title mentions multi-million year evolution but we do not see any results for multi-million year dynamics, or explore the effects of changing ocean basins over deep time on the model chemistry. But presumably these tests will come in future papers.

I found the following corrections or questions when reading through:

Line 6: suggest to give an idea of the timescales here

Line 35: suggest “nondimensional” rather than “low-resolution”

Line 106: typo “u-inside”

Line 127 and 131: both should be “deep” rather than “surface”?

Line 164: “sulw” should be “silw”?

Line 174 and 194: “Ffocw” - organic carbon weathering – is not included in the CO2 mass balance but should produce CO2? Some further explanation of the oxygen cycle mass balance would also be useful, e.g. how it differs from the GEOCARBSULF.

Line 295: “This 10-fold reduction in coastal boxes was tuned in order to avoid massive precipitation of carbonates in coastal surface boxes”. Does this hint that the formulation is not very accurate in the first place? Why is the initial prediction so much higher than in the real world?

Line 557: hydrothermal burial is with iron oxides?

Line 617 “abundance”?

Line 698. But what does “E” represent here?

Line 939 “to reproduce” typo

Line 978: suggest replace GEOCLIM’s with GEOCLIM

Figure 4: it would be nice to have labels on the subplots saying which tracer they are

Line 1007: suggest “found it necessary”

Line 1010: suggest replace “free” with “behave freely”

Line 1122: repetition of the model names could be removed here

Line 1151: suggest replace “way more complex” with “significantly more complex” or similar

Figure 8: titles or labels showing what is in each panel would be very helpful here.

Figure 11: as above, labels on y axes or elsewhere would make this much easier to read