Background: The nucleotide excision repair (NER) pathway participates in platinum-induced DNA damage repair. Single nucleotide polymorphisms (SNPs) in miRNA-binding sites in the NER genes RPA2 and GTF2H1 are associated with the risk of colorectal cancer (CRC). Here, we analyzed whether RPA2 and GTF2H1 SNPs predict the efficacy of oxaliplatin in metastatic CRC (mCRC) patients. Patients and methods: Genomic DNA was extracted from blood samples from 457 patients with mCRC enrolled in the TRIBE trial, which compared first-line FOLFOXIRI plus bevacizumab (BEV) (n = 230, discovery cohort) and first-line FOLFIRI plus BEV (n = 227, control cohort). SNPs were analyzed by PCR-based direct sequencing. Results: In the FOLFOXIRI + BEV-treated cohort expressing wild-type KRAS, progression-free survival (PFS) was shorter for the RPA2 rs7356 C/C variant subgroup than the any T allele subgroup in univariate analysis (9.1 versus 13.3 months respectively, hazard ratio (HR) 2.32, 95% confidence interval (CI): 1.07–5.03, p = 0.020) and this remained significant in multivariable analysis (HR 2.97, 95%CI: 1.27–6.94, p = 0.012). A similar trend was observed for overall survival. In contrast, patients expressing mutant RAS and RPA2 rs7356 C/C variant had longer PFS with FOLFOXIRI + BEV than with FOLFIRI + BEV (12.1 versus 7.6 months, HR 0.23, 95%CI: 0.09–0.62, p = 0.002) but no superiority of FOLFOXIRI + BEV was observed for the RAS mutant, RPA2 rs7356 any T variant subgroup (11.7 versus 9.6 months, HR 0.77, 95%CI: 0.56–1.07, p = 0.12) or the RAS wild-type, RPA2 rs7356 C/C variant subgroup. Conclusion: RPA2 SNPs may serve as predictive and prognostic markers of oxaliplatin responsiveness in a RAS status-dependent manner in mCRC patients receiving FOLFOXIRI + BEV.