The Gulf of Mexico circulation is examined from the results of a high-resolution (1/12¿) North Atlantic simulation using the Miami Isopycnic Coordinate Ocean Model. The motivation for this paper is twofold: first, we validate the model's performance in the Gulf of Mexico by comparing the model fields to past and recent observations, and second, given the good agreement with the observed Gulf of Mexico surface circulation and Loop Current variability, we expand the discussion and analysis of the model circulation to areas that have not been extensively observed/analyzed, such as the vertical structure of the Loop Current and associated eddies, especially the deep circulation below 1500 m. The interval between successive model eddy sheddings is 3 to 15 months, the eddy diameters range between 140 and 500 km, the life span is about 1 year, and the translational speeds are 2--3 km d-1, in good agreement with observations. Areas of high cyclonic eddy occurrence in the model are southwest of Florida, the Loop Current boundary, and the western Campeche Bay area. The cyclonic eddy diameters range between 50 and 375 km, the orbital speeds range between 1 and 55 cm s-1, the translational speeds range between 0.5 and 14 km d-1, and the eddy life spans range between 1 and 3 months. The vertical structure of the temperature and salinity of each modeled eddy, from the moment it is shed until it disintegrates in the western Gulf of Mexico, is in agreement with the few available observations. Below 1500 m, deep cyclonic eddies are associated with the surface Loop Current anticyclones. The eddy variability is consistent with Rossby waves propagating westward, and there is bottom intensification of the flow close to steep topography. Overall, we show that this very high horizontal resolution isopycnic coordinate ocean model, which is able to produce a quite realistic surface circulation for the North and equatorial Atlantic, is also able to reproduce well the smaller-scale, basin-wide intricate dynamics such as the Gulf of Mexico variability.