The Energy Exascale Earth System Model (E3SM) is a new fully coupled model that includes sea ice and ocean components on an unstructured mesh. This means that high resolution may be applied to regions of the simulated earth to resolve oceanographic phenomena important for aspects of the coupled system, while preserving computationally efficiency elsewhere. This new modeling approach is particularly advantageous for investigating polar regions, where highly-resolved horizontal gradients in the saline-stratified oceans result in improved vertical mixing, while sea ice circulation is more accurately influenced by bathymetry according to Sverdrup’s relation. In a series of fully coupled simulations, we have explored the impact of resolution and regional focus of the E3SM ice-ocean mesh, with and without enhanced atmosphere resolution. We compare sea ice results from our standard resolution 30-60km oceanic mesh, 6-18km global high resolution simulation, and a regionally-focused Labrador Sea mesh with 10-60km resolution globally. The latter grid configuration only enhances resolution in the Bering Strait, Canadian Archipelago and around Greenland, and includes a resolution transition to non-eddy permitting cells in the midst of the Arctic Ocean. This grid was chosen to demonstrate the capability of E3SM to improve a Labrador Sea ice extent bias in the E3SM Coupled Model Inter-comparison Project (CMIP6) submissions, which it has largely accomplished for less than 10% extra computational cost over the background 30-60km mesh, all with the same atmospheric resolution. However, we pay the price of including a resolution transition across the middle of the Arctic Ocean, which degrades deformation characteristics of the ice, and the integrated ice-ocean barotropic mode in the Canada and Eurasian Basins. This does not occur in the global high resolution 6-18km eddy-permitting simulations. Using new data fusion techniques including an altimetric satellite emulator, we demonstrate the internal variability in sea ice volume in resolution dependent, which we attribute to better resolved ocean currents, atmospheric coupling, and sea ice deformation. This has important consequences for understanding and simulating internal polar variability in the coupled Earth System.