Neoadjuvant chemotherapy (NACT) prior to liver resection is advantageous for patients with colorectal cancer liver metastases (CLM). Bevacizumab- or cetuximab-based NACT may affect patient outcome and curative resection rate, but comparative studies on differential tumour regression grade (TRG) associated with distinct antibodies-associated regimens are lacking. Ninety-three consecutive patients received NACT plus bevacizumab (n = 46) or cetuximab (n = 47) followed by CLM resection. Pathological response was determined in each resected metastasis as TRG rated from 1 (complete) to 5 (no response). Except for KRAS mutations prevailing in bevacizumab versus cetuximab (57 vs. 21 %, p = 0.001), patients characteristics were well balanced. Median follow-up was 31 months (IQR 17-48). Bevacizumab induced significantly better pathological response rates (TRG1-3: 78 vs. 34 %, p < 0.001) as well as complete responses (TRG1: 13 vs. 0 %, p = 0.012) with respect to cetuximab. Three-year progression-free survival (PFS) and overall survival (OS) were not significantly different in the two cohorts. At multivariable analysis, significant association with pathological response was found for number of resected metastases (p = 0.015) and bevacizumab allocation (p < 0.001), while KRAS mutation showed only a trend. Significant association with poorer PFS and OS was found for low grades of pathological response (p = 0.009 and p < 0.001, respectively), R2 resection or presence of extrahepatic disease (both p < 0.001) and presence of KRAS mutation (p = 0.007 and p < 0.001, respectively). Bevacizumab-based regimens, although influenced by the number of metastases and KRAS status, improve significantly pathological response if compared to cetuximab-based NACT. Possible differential impact among regimens on patient outcome has still to be elucidated.