PURPOSE: Clinical outcome of TKA remains unsatisfactory in 20% of the cases. Navigation has added accuracy in terms of alignment, but has improved clinical outcome only in small series with gap-balanced techniques. Reason for that could be that conventional gap balanced TKA determines gaps in extension and 90° of flexion only. Furthermore, measurement is only static. Therefore, the accuracy of a new dynamic navigation software which allows gap assessment throughout the entire range of motion was tested. The purpose of this study was to investigate the accuracy and reliability of dynamic gap testing during gap-balanced TKA. METHODS: In two different centres, a total of 65 TKA procedures were performed in a tibia-first, gap-balanced technique using a new CAS software. At the same and at different time points of surgery, two different surgeons performed gap measurement to provide inter-observer reliability data and repeated gap measurement to provide intra-observer reliability data. These gap measurements were performed throughout the entire ROM under dynamic stress testing to detect maximum gap values. RESULTS: CAS surgery was able to produce correct coronal alignment in 96.4% of the cases (within 3° mechanical alignment). Both inter-observer and intra-observer reliabilities were excellent for gap values throughout the entire ROM. Inter-observer bias of deviation 0.05; 95% limits of agreement of - 2.1 to + 2.21 mm. Intra-observer bias of deviation 0.09; 95% limits of agreement of - 2.27 to + 2.44 mm. CONCLUSIONS: This new CAS software in combination with the presented dynamic gap measurement provides accurate gap values and therefore facilitates balancing TKA. This technique works reproducibly for different surgeons and has proven robustness also for repeated measurements of any surgeon in this study.