This study aimed (I) to compare the number of repetitions that can be completed to failure (XRM) and before reaching a 15%, 30%, or 45% velocity loss threshold (XVLT) in the bench press exercise performed using different grip widths, and (II) to examine the inter-individual variability in the percentage of completed repetitions with respect to the XRM when the set volume is prescribed based on a fixed number of repetitions (FNR) and several velocity loss thresholds (VLT). Nineteen men performed four separate sessions in a random order where there was a single set of repetitions completed to failure against 75% of the one-repetition maximum during the Smith machine bench press exercise using a narrow, medium, wide, or self-selected grip widths. The XRM ( = 0.545) and XVLTs ( ≥ 0.682) were not significantly affected by grip width. A high and comparable inter-individual variability in the percentage of completed repetitions with respect to the XRM was observed when using both an FNR (median CV = 24.3%) and VLTs (median CV = 23.5%). These results indicate that Smith machine bench press training volume is not influenced by the grip width and that VLTs do not allow a more homogeneous prescription of the set volume with respect to the XRM than the traditional FNR.