Full recycling of transuranic (TRU) isotopes can in theory lead to a reduction in repository radiotoxicity to reference levels in as little as ∼500 years provided reprocessing and fuel fabrication losses are limited. However, over a limited timeframe, the radiotoxicity of the ‘final’ core can dominate over reprocessing losses, leading to a much lower reduction in radiotoxicity compared to that achievable at equilibrium. In Part I of this paper, TRU recycle over up to 5 generations of light water reactors (LWRs) or sodium-cooled fast reactors (SFRs) is considered for uranium (U) fuel cycles. With full actinide recycling, at least 6 generations of SFRs are required in a gradual phase-out of nuclear power to achieve transmutation performance approaching the theoretical equilibrium performance. U-fuelled SFRs operating a break-even fuel cycle are not particularly effective at reducing repository radiotoxicity as the final core load dominates over a very long timeframe. In this paper, the analysis is extended to the thorium (Th) fuel cycle. Closed Th-based fuel cycles are well known to have lower equilibrium radiotoxicity than U-based fuel cycles but the time taken to reach equilibrium is generally very long. Th burner fuel cycles with SFRs are found to result in very similar radiotoxicity to U burner fuel cycles with SFRs for one less generation of reactors, provided that protactinium (Pa) is recycled. Th-fuelled reduced-moderation boiling water reactors (RBWRs) are also considered, but for burner fuel cycles their performance is substantially worse, with the waste taking ∼3–5 times longer to decay to the reference level than for Th-fuelled SFRs with the same number of generations. Th break-even fuel cycles require ∼3 generations of operation before their waste radiotoxicity benefits result in decay to the reference level in ∼1000 years. While this is a very long timeframe, it is roughly half that required for waste from the Th or U burner fuel cycle to decay to the reference level, and less than a tenth that required for the U break-even fuel cycle. The improved performance over burner fuel cycles is due to a more substantial contribution of energy generated by 233U leading to lower radiotoxicity per unit energy generation. To some extent this an argument based on how the radiotoxicity is normalised: operating a break-even fuel cycle rather than phasing out nuclear power using a burner fuel cycle results in higher repository radiotoxicity in absolute terms. The advantage of Th break-even fuel cycles is also contingent on recycling Pa, and reprocessing losses are significant also for a small number of generations due to the need to effectively burn down the TRU. The integrated decay heat over the scenario timeframe is almost twice as high for a break-even Th fuel cycle than a break-even U fuel cycle when using SFRs, as a result of much higher 90Sr production, which subsequently decays into 90Y. The peak decay heat is comparable. As decay heat at vitrification and repository decay heat affect repository sizing, this may weaken the argument for the Th cycle.