Third-generation concentrating solar-thermal power (CSP) systems are proposed which aim to halve the cost of electricity from CSP. This work demonstrates that a transition from state-of-the-art cylindrical molten nitrate salt receivers with a peak concentrated solar flux of 1MW/m2 and a heliostat field optical error of 3mrad to future cylindrical sodium receivers with a peak flux of 1.8MW/m2 and a heliostat field optical error of 1mrad could enable the exergy output of CSP receivers to be increased significantly, by 25% at design point, while keeping the general system design very similar. This comparison is conducted on a large dataset established with a detailed field and receiver energy balance model considering field aiming, peak receiver flux and hydrodynamics. It is based on design point simulations, and is independent from cost considerations. At high temperatures, the optimal configurations achieve their gains only with very accurate heliostats and when the peak flux is increased, while at low temperatures, there is nothing to be gained from very high peak flux limits on external tubular receivers. This underscores the critical importance of high heliostat accuracy for high-temperature CSP systems design. For the third-generation, high-temperature (740 °C) sodium receiver configuration considered, it is found that the upper limit to thermal–optical efficiency of optimised receivers at design-point is around 88 %.