The effect of herbicide rates on weed control and crop yield is the subject of countless and ongoing research projects. Weed seed banks receive very little attention in comparison. The seed bank resulting from 3 yr (2006 to 2008) of single herbicide rates in a cropping system where glyphosate/glufosinate and corn/soybean were rotated or not was evaluated in a field located in St-Augustin-de-Desmaures, Québec, Canada. Field plots under conventional tillage were seeded in corn every year, or corn and soybean (1 yr). These plots received the same herbicide every year or various glyphosate/glufosinate 3-yr sequences. Subplots were sprayed with a single POST application of the recommended rate of glyphosate (900 g ae ha−1) or glufosinate (500 g ai ha−1) or lower rates. Subplots received the same full (1.0×, recommended) or reduced (0.5×, 0.75×) rate every year. After crop harvest in 2008, soil cores were extracted and the weed seed bank was evaluated. Including soybean in the cropping system resulted in lower seed banks compared to those under continuous corn cropping. Including glufosinate in a glyphosate herbicide sequence increased weed seed banks due to the lower efficacy of the glufosinate rates tested at reducing the seed bank of annual grasses. Higher herbicide rates translated into lower seed banks, up to a certain rate. After 3 yr, the lowest seed bank (full glyphosate rates every year) still had 4,339 ± 836 seeds m−2 and was higher than the initial seed bank (2,826 ± 724 seeds m−2).