For banding-free bSSFP cardiac cine, a highly-accelerated projection-reconstruction sequence acquires three phase-cycles within a short breathhold. Data is also acquired on the rewinds, enabling generation of B0 maps using BMART, which are used for phase-cycle combination. We show that this data is well-captured by a multi-scale low-rank (MSLR) model, which recovers the normal and rewind images from the aggressively-undersampled data with less streaking and blurring than total-variation-regularized ESPIRiT. In addition to improving the phase-cycle component images, MSLR facilitates generation of temporally-resolved B0 maps with good SNR. Together, these two improvements result in the final, field-map-combined cine images having high quality.