This is an important and timely piece of work, particularly considering the next phase of MioMIP. The methodology is sound and is generally well described. I would like to see a number of minor comments addressed before publication and attach a PDF which details these. Many comments relate to rephrasing or editorial-style changes that will enhance the paper for the reader. 

Generally, the paper would benefit from more discussion around the implications of the findings. As there is no specific 'Discussion' section, I would like to see this weaved into all other results sections. I have pointed out some specific instances in the PDF.

I would also like to see some more nuance when it comes to comparing the model results to the proxy reconstructions. As it stands, the text largely reads that proxy data are correct and the error must all come from the models, when it is very likely a combination of both model error and proxy bias. We need to better understand both sources of error, and we shouldn't be adapting the model CO₂ to potentially unlikely/unevidenced levels just to better represent the proxy values. Again, I have pointed out some specific instances in the PDF. 

Figures are generally very good quality, but some would benefit from subplot labels (a, b, c etc.) so that they can be clearly identified in the caption and the main text.

Please also check for the consistency in subscripting CO₂. 

Comments to the Author:

This manuscript presents a comprehensive sensitivity study of the mid-Miocene climate using the CESM1.2 model. The study aims to understand why climate models systematically underestimate high-latitude warmth during the Mid-Miocene Climatic Optimum and to quantify equilibrium climate sensitivity (ECS) from this warm paleoclimate. The study explores the roles of paleogeography, CO₂ concentration, atmospheric model physics, solar forcing, and Antarctic ice sheet. The topic is highly relevant, given the mid-Miocene's potential as an analog for near-future warm climates and the persistent model-data mismatch, especially at high latitudes. Here, I give some comments.

Major Comments:

Paleogeographic Reconstruction:

The study relies on an unpublished paleogeography from Getech Plc., which is compared with the Burls et al. (2021) reconstruction. While this offers a valuable comparison, the manuscript provides insufficient details on the paleogeographic reconstructions, and does not compare with several recent published mid-Miocene paleogeographic reconstructions. As the paleogeography is the important boundary condition that fundamentally shapes the simulated climate.

Model-Data Mismatch:

The manuscript highlights the persistent cold bias in high latitudes. It is advisable to consider adding a discussion on "seasonal deviation" of the proxy. And the discussion could be strengthened by linking this to broader challenges in paleoclimate modeling (e.g., cloud feedbacks, ocean heat transport, vegetation-atmosphere interactions).

Addition, a key concern is that the model might be over-reliant on CO₂ forcing to achieve a warmer climate, potentially at the expense of other, less constrained factors that are known to be important for warm paleoclimates, such as the role of ocean–ice interactions and their dependence on model dynamics.

The simulated AMOC:

The model simulation suggests a collapse of the Atlantic Meridional Overturning Circulation (AMOC) during the middle Miocene. However, the available geological evidence does not support a complete shutdown of the AMOC at that time.

Antarctic Ice Sheet Impact:

The finding that the inclusion of an Antarctic ice sheet leads to global warming is intriguing and counterintuitive. While the authors briefly discuss this, the mechanism remains speculative. Provide the relative contribution percentages of each feedback item (cloud, water vapor, albedo, temperature) to the 1.4 °C increase. This could be a significant contribution if better explained.