In this manuscript, the authors have comprehensively shown the results of a multi-year ICON simulation with a NWP configuration and a 2.5-km horizontal mesh. The analyses focused on mean precipitation and radiative fields and global statistics of local-to-synoptic-scale weather phenomena such as MCSs, tropical cyclones, and equatorial waves, compared with several observational data sets. The evaluation of these aspects suggested that complied statistics of local winds and precipitation as well as the mean precipitation distribution are reproduced relatively well, whereas the simulation still has inevitable biases, as found in the morphology of MCSs, tropical waves (incl. the MJO), and radiative budgets. 

I acknowledge that this is the first step of addressing weather and climate using this version of ICON, and that it is worth being reported as one of milestones. Meanwhile, some of the presented results are interpreted speculatively without concrete evidence. Also, I wonder what clear merits of this ICON simulation are, compared to preexisting km-scale regional climate simulations, because the results except for global mean fields weigh heavy on regional statistics of meso-to-synoptic-scale variability (rather than large-scale circulation), which could have been addressed by regional modeling.  Furthermore, I feel that the Introduction does not precisely describe historical advances in global km-scale modeling, with too much emphasis on the recent trend. Based on these points, I think this paper must undergo major revision before it can be considered for publication.

[Major comments]

1. The authors have tried to interpret a source of the biases in several subsections: for example, reasons for high temperature biases over land, for lower differences at high wind speeds over several regions, and for underestimated equatorial waves. However, I wonder if their interpretation could be done by relatively narrow insights without sufficient evidence. While I do not intend to request very detailed analyses, it would be better to provide more hints about the emergence of biases for the future model improvement. The suggestions and/or issues are listed below.

* Impacts of cloud radiative forcing, in addition to the issues of land-atmosphere coupling, on the surface temperature bias (cf. Sections 3.1 and 3.2)

* Representation of extratropical cyclones around CNA, WCE, and eastern North America (cf. LL. 307-310)

* Equatorial Rossby waves are affected by coupling between moisture and dynamics (e.g., Yasunaga and Mapes, 2012. JAS; Yasunaga et al., 2019, JCLI; Nakamura and Takayabu, 2022, JAS), not just by dynamics. 

2. While the authors presented the geographical variability of mesoscale phenomena (e.g., MCSs, diurnal cycles) in Figures 5, 6, 7, 11, and 12, I wonder how different it is simulated when comparing the accumulated results from the regional km-scale modeling framework. It is true that the presented results have global aspects, but they should be somewhat described without global models. I would appreciate it if the authors could discuss added values of this study to address the above points.

3. The descriptions about the advances in global km-scale modeling are heavily biased by the recent trend observed in Europe. This kind of activities started two decades ago in the Japanese community with the Nonhydrostatic ICosahedral Atmospheric Model (NICAM), and it has published many research articles describing its importance of both weather and climate modeling. To ensure the correct historical advances in science, please reorganize the Introduction with appropriate citations (please see also the specific comments). 

[Specific and/or minor comments]

Title: I feel that the title is exaggerated compared to the contents in the main text. What are "global waves", despite not fully mentioning planetary-scale waves? I would like the authors to reconsider the title to be consistent with the fact that the present study mainly addresses global statistics of meso-to-synoptic-scale features. 

LL.13-14: I do not fully agree on this speculation. I wonder if the poor representation of convectively coupled equatorial waves is attributed to misrepresentation of thermodynamic-convection coupling (cf., Takasuka, Becker, and Bao, 2025).

LL.14-15: I wonder if the main text, especially the concluding section, does not provide the sufficient value of multi-year global km-scale simulations, even though it tells us some information about biases. 

LL.25-29: "historically restricted its use to..." has been true in the main stream, but NICAM already succeeded in this type of simulation. I would like the authors to reflect the historical review by Satoh et al. (2019) in the revised Introduction. 

LL.36-37: Sato et al. (2009, JCLI) also showed the better representation of diurnal cycles in the global km-scale model. 

LL.37-39: Miura et al. (2007, Science) showed the first success of the realistic MJO simulation and associated tropical cyclogenesis, featured by the realistically simulated multi-scale convective organization.

LL.49-50: The same comment as that for LL.25-29. 

L.60: These were already (i.e., before 2020s) shown by Miura et al. (2007, Science), Nasuno et al. (2008, JAS), Holloway et al. (2012, JAS)...

LL.65-66: Miura et al. (2023, BAMS) have also pointed out this direction. 

L.129: slow-evolving waves -> slow-evolving variability (because the MJO is not a dynamical wave...)

LL.228-232: In addition to this problem, it seems that a large bias of radiation budget (at the TOA and SFC) has impacts on T2m bias. Also, how are the distributions of cloud radiative forcing?

LL.259-260: This does not necessarily hold true for all the regions; for example, over the North America and South America in the subtropics. 

LL.266-271: Figure 2a shows the weak precipitation band in the Southern Hemisphere very near the equator. I think this can be a glimpse of a double ITCZ... Also, how about mentioning the reproducibility of precipitation bands in the mid-latitude?

L.271: Note that Takasuka et al. (2024, JAMES) showed that resolving the double ITCZ problem was achieved by the reconsideration of microphysics. 

LL.291-292: I'm a bit surprised at this, because the phase lag of diurnal cycles of precipitation has been found in IMERG, which uses passive sensors for the estimation of precipitation, compared to radar-based precipitation products. 

LL.321-322: What is a possible reason for the low bias over the North Atlantic Basin? I wonder if this could be related to the underestimation of easterly waves (similar to equatorial waves, as shown in Fig. 9)

LL.327-328: Is this attributed to the poor representation of rapid intensification? Also, Baker et al. (2024, GRL) should be cited somewhere in this paragraph, because they already showed benefits of km-scale models in representing tropical cyclones.

L.322: Equatorial waves -> Equatorial waves and the Madden-Julian oscillation (because MJO is not a equatorial wave.)

L.339, L.341: Please cite appropriate references. 

Section 3.7: It would be better to provide a brief description about the criteria for the detection of MCSs. Readers may not read other paper carefully that introduces the methodology. 

LL.359-360: The same comment as that for LL.13-14.

LL.400-401: Is this also related to underestimated shortwave incoming (especially over ocean)? I wonder if water clouds have higher albedo and thus prevent radiation from reaching the surface.