This study compares the diagnostic performance of a second generation anti-cyclic citrullinated peptide antibody (CCP2) with a third generation anti-CCP antibodies assay (CCP3), as well as the combination of both tests. Serum samples of 127 patients were analyzed. IgG anti-CCP 2 and IgM rheumatoid factor were determined by EliA™ technique on a Phadia 250 instrument (Thermo Fisher Scientific), anti-CCP3 by the Quanta Flash™ anti-CCP3 IgG kit, BIO-FLASH Rapid Response Chemiluminscence Analyzer (INOVA Diagnostics). Diagnostic performance was compared using ROC-curves, sensitivity, specificity, likelihood ratios, and predictive values. Logistic regressions were used to investigate whether using both tests (anti-CCP2 and anti-CCP3) gives a better prediction of rheumatoid arthritis. At the manufacturer’s cut-offs sensitivity and specificity were 79.4 and 61.0% for CCP3 and 80.9 and 69.5% for CCP2. No significant differences could be observed regarding the areas under the curve (AUC) of both ROC-curves. The optimal cut-off point for CCP2 was 10.5 U/ml (sensitivity of 75.0% and specificity of 80.0%) and 5.6 U/ml for CCP3 (sensitivity of 86.9% and specificity of 61.0%). Binary logistic regressions indicated that the likelihood of having rheumatoid arthritis (RA) is significantly higher when testing positive on both CCP2 and CCP3 compared to CCP2 or CCP3 alone. In our cohort, comparable performance was found between the two CCP assays. Positivity for both CCP2 and CCP3 resulted in the most specific identification of RA patients. In patients with joint complaints suspected of having RA and with a weakly positive CCP 2 (≥7 and ≤16 U/ml) CCP3 testing could be of additive value for diagnosing RA.