Since the earlier liquid-metal fast breeder reactor transient overpower assessments were done (1975), new experimental data and modeling improvements have occurred that indicate later failures and more molten fuel squirted into the channel with a higher propensity for plugging. An initial sweepout still occurs, and an analysis shows that even if coherent instead of the expected stochastic failures occur, the blockages are partial, the reactor is strongly shut down, and a coolable geometry exists. Hence, the overall consequences would be benign.