To study the complexity of hot Jupiter atmospheres revealed by observations of increasing quality, we have adapted the UK Met Office Global Circulation Model (GCM), the Unified Model (UM), to these exoplanets. The UM solves the full 3D Navier-Stokes equations with a height-varying gravity, avoiding the simplifications used in most GCMs currently applied to exoplanets. In this work we present the coupling of the UM dynamical core to an accurate radiation scheme based on the two-stream approximation and correlated-k method with state-of-the-art opacities from ExoMol. Our first application of this model is devoted to the extensively studied hot Jupiter HD 209458b. We have derived synthetic emission spectra and phase curves, and compare them to both previous models also based on state-of-the-art radiative transfer, and to observations. We find a reasonable agreement between observations and both our days side emission and hot spot offset, however, our night side emissions is too large. Overall our results are qualitatively similar to those found by Showman et al. (2009, ApJ, 699, 564) with the SPARC/MITgcm, however, we note several quantitative differences: Our simulations show significant variation in the position of the hottest part of the atmosphere with pressure, as expected from simple timescale arguments, and in contrast to the “vertical coherency” found by Showman et al. (2009). We also see significant quantitative differences in calculated synthetic observations. Our comparisons strengthen the need for detailed intercomparisons of dynamical cores, radiation schemes and post-processing tools to understand these differences. This effort is necessary in order to make robust conclusions about these atmospheres based on GCM results.