Fission of fissile uranium or plutonium nucleus in nuclear fuel results in fission products. A small fraction of them are volatile and can migrate under the effect of concentration gradients to the grain boundaries of the fuel pellet. Eventually, some fission gases are released to the rod void volumes by a thermally activated process. Local transients of power generation could distort even further the already non-uniform axial power and fission gas concentration profiles in fuel rods. Most of the current fuel rod performance codes neglects these gradients and the resulting axial fission gas transport (i.e., gas mixing is considered instantaneous). Experimental evidences, however, highlight axial gas mixing as a real time-dependent process. The thermal feedback between fission gas release, gap composition and fuel temperature, make the “prompt mixing assumption” in fuel performance codes a key point to investigate due to its potential safety implications. This paper discusses the possible scenarios where axial transport can become significant. Once the scenarios are well characterized, the available database is explored and the reported models are reviewed to highlight their major advantages and shortcomings. The convection-diffusion approach is adopted to simulate the axial transport by decoupling both motion mechanisms (i.e., convection transport assumed to be instantaneous) and a stand-alone code has been developed. By using this code together with FRAPCON-3, a prospective calculation of the potential impact of axial mixing is conducted. The results show that under specific but feasible conditions, the assumption of “prompt axial mixing” could result in temperature underestimates for long periods of time. Given the coupling between fuel rod thermal state and fission gas release to the gap, fuel performance codes predictions could deviate non-conservatively. This work is framed within the CSN-CIEMAT agreement on “Thermo-Mechanical Behaviour of the Nuclear Fuel at High Burnup”.