The present paper is related to the design and neutronic characterization of the principal control assembly system for the reference large (2400 MWth) Generation IV gas-cooled fast reactor (GFR), which makes use of ceramic–ceramic (CERCER) plate-type fuel-elements with (U–Pu) carbide fuel contained within a SiC inert matrix. For the neutronic calculations, the deterministic code system ERANOS-2.0 has been used, in association with a full core model including a European fast reactor (EFR)-type pattern for the control assemblies as a starting point. More specifically, the core contains a total of 33 control (control system device: CSD) and safety (diverse safety device: DSD) assemblies implemented in three banks. In the design of the new control assembly system, particular attention was given to the heat generation within the assemblies, so that both neutronic and thermal–hydraulic constraints could be appropriately accounted for. The thermal–hydraulic calculations have been performed with the code COPERNIC, significant coolant mass flow rates being found necessary to maintain acceptable cladding temperatures of the absorber pins. Complementary to the design study, neutronic investigations have been performed to assess the impact of the control assemblies in the GFR core in greater detail (rod interactions, shift of the flux, peaking factors, etc.). Thus, considerable shadowing effects have been observed between the first bank and the safety bank, as also between individual assemblies within the first bank. Large anti-shadowing effects also occur, the most prominent being that between the two CSD banks, where the total assembly worth is almost doubled in comparison to the sum of the individual values. Additional investigations have been performed and, in this context, it has been found that computation of the first-order eigenvalue and the eigenvalue separation is a robust tool to anticipate control assembly interactions in a large fast-spectrum core. One interesting finding is that the interactions are lower when one of the control assembly banks is located at a radial position corresponding to the minimum in the first-order eigenvalue (at approximately half the core radius). An important aspect of the design optimization process has been to obtain a low heterogeneity effect for the control assembly worth. This was achieved by applying a methodology based on reactivity equivalence, with the optimization goal set at minimizing the shadowing effects between the absorber pins as well as the self-shielding within the absorber pins. A relatively low reduction of aproximatively 13percent was obtained as the heterogeneity effect for the control assembly worth, the corresponding value in the case of Superphénix being significantly higher, viz. about 20 percent.