This reviewer finds the manuscript to be of high technical value, well organized, and enjoyable to read.  The methods employed in understanding the relative strengths and weaknesses of various wind turbine aerodynamic modeling approaches, particularly for floating wind turbine applications, are sound.  The results provided are comprehensive and cover a wide range and large quantity of possible numerical modeling approaches.  The discussions surrounding the results are insightful, and the conclusions provide useful information for future offshore wind turbine designers and modelers (i.e., undsteady aerodynamic models will be required for cases employing realistic generator and blade pitch controls).  This reviewer only found a few minor technical issues which can be easily remedied with a proofread (e.g., line 157 should probably read 'However, in reality it takes time...', line 159 should likely read '...(also referred to as dynamic wake)..., etc.).

The paper presents the results of the benchmark exercise OC6 Phase III. The main focus of this joint venture is to compare predictions of aerodynamic models of varying fidelity among each other but also against wind tunnel measured datasets, for a rotor undergoing surge and pitching motion, that emulates wave induced motions of a FOWT.

The work presented in the paper constitutes a huge effort, both in terms of computational and human resources. The overall agreement of several models of varying fidelity is remarkable in many of the addressed conditions while differences, wherever they are observed, are plausible given the range of fidelity and the underlying assumptions in the different models. This agreement also implies a substantial effort by the coordinator of the task, to bridge the uncertainties due to inevitable misinterpretations of numerical and physical parameters that always occur in such extended benchmark exercises. Furthermore the knowledge gained by this exercise will be definitely valuable to the scientific community of rotor aerodynamicists and floating offshore technology.

Given the above, I believe that the paper is suitable for publication in WES journal. However, I would recommend the authors to revise their document, taking into account all my comments in the accompanying pdf. The most important ones are summarized below:

1) The authors state in page 7 line 161 “In GDW dynamic inflow is explicitly calculated through the use of an apparent mass in the induction calculation”. Traditionally, apparent mass in aerodynamics is the unsteady part of the pressure loads and it is not connected to hysteresis due to time varying wake induction (which is what GDW does). I would describe GDW as a semi-analytical frozen wake variant of a vortex wake model.

2) FVW models are further categorized to lifting line, lifting surface and 3D panel. Not all of them use pre-calculated polars and this is an important distinction that should be incorporated in table 4.

3) In asymmetric flow conditions, 2P, 4P etc. harmonics appear in the loads of the blade rotating frame (system attached to the rotating blade) which become 3P, 6P etc. when transferred to the tower frame. These are not only due to tower shadow. So, by low pass filtering you may omit part of the aerodynamic unsteady response. A band stop filter would be preferable in this case. Unless you are sure that loads are described by a pure 1P response which I doubt in the case of CFD or FWM. Or of course, if you are only interested in resolving the dynamic inflow effects due to the low frequency platform motion. In this case a proper explanation is required though.

4) I would recommend the authors to remove the FWM results for the wake inflection point from figure 9. The answer why these results are pointless can be found in the explanation by the authors. It is also necessary to provide more information on how much is the ambient turbulence in the wind tunnel test data and how this is simulated by the CFD tools.

5) There are several small syntax and grammar errors which however do not interfere with the understanding of the material. I corrected as many as I could in the accompanying pdf but I’m quite convinced that I missed many. Therefore I would recommend the authors to proof read the text before resubmitting it.