The latitudinal and temporal variation of atomic oxygen (O) is opposite between the empirical model, MSIS and the whole atmosphere model, WACCM-X at 97-100 km. The [O] from WACCM-X has maxima at solstices and summer mid-high latitudes, similar to [O] from SABER. We use the densities and dynamics from WACCM-X to drive the Global Ionosphere Thermosphere Model (GITM) at its lower boundary, and compare it with the MSIS driven GITM. We focus on the differences in the modeling of the thermospheric and ionospheric semiannual oscillation (T-I SAO). Our results reveal that driving GITM with WACCM-X shifts the phase of T-I SAO to maximize around solstices. Nudging the dynamics in GITM towards WACCM-X, reduces the amplitude of the oppositely-phased SAO but does not completely correct its phase. We find that during solstices, WACCM-X driven GITM has a smaller temperature gradient between the hemispheres and weaker meridional and vertical winds in the summer hemisphere. This leads to accumulation of [O] at lower latitudes due to weaker meridional transport, resulting in solstitial maxima in global means. WACCM-X itself has the right phase of SAO in the upper thermosphere but wrong at lower altitudes. The exact mechanisms that can correct the phase of SAO in IT models while using SABER-like [O] in the MLT are currently unknown and warrant further investigation. We suggest mechanisms that can reduce the solstitial maxima in the lower thermosphere, for example, stronger interhemispheric meridional winds, stronger residual circulation, seasonal variation in eddy diffusion, and momentum from breaking gravity waves.