We aimed to explore SARS-CoV-2 evolution during in vitro neutralisation using next generation sequencing, and to determine whether sera from individuals immunised with two doses of the Pfizer-BioNTech vaccine (BNT162b2) were as effective at neutralising the variant of concern (VOC) Delta (B.1.617.2) compared to the earlier lineages Beta (B.1.351) and wild-type (A.2.2) virus. Using a live-virus SARS-CoV-2 neutralisation assay in Vero E6 cells, we determined neutralising antibody titres (nAbT) against three SARS-CoV-2 strains (wild type, Beta, and Delta) in 14 participants (vaccine-naïve (n = 2) and post-second dose of BNT162b2 vaccination (n = 12)), median age 45 years [IQR 29–65]; the median time after the second dose was 21 days [IQR 19–28]. The determination of nAbT was based on cytopathic effect (CPE) and in-house quantitative reverse transcriptase real-time quantitative polymerase chain reaction (RT-qPCR) to confirm SARS-CoV-2 replication. A total of 110 representative samples including inoculum, neutralisation breakpoints at 72 h, and negative and positive controls underwent genome sequencing. By integrating live-virus neutralisation assays with deep sequencing, we characterised both functional antibody responses and accompanying viral genetic changes. There was a reduction in nAbT observed against the Delta and Beta VOC compared with wild type, 4.4-fold (p ≤ 0.0006) and 2.3-fold (p = 0.0140), respectively. Neutralising antibodies were not detected in one vaccinated immunosuppressed participant and the vaccine-naïve participants (n = 2). The highest nAbT against the SARS-CoV-2 variants investigated was obtained from a participant who was vaccinated following SARS-CoV-2 infection 12 months prior. Limited consensus level mutations occurred in the various SARS-CoV-2 lineage genomes during in vitro neutralisation; however, consistent minority allele frequency variants (MFV) were detected in the SARS-CoV-2 polypeptide, spike (S), and membrane protein. Findings from countries with high COVID-19 incidence may not be applicable to low-incidence settings such as Australia; as seen in our cohort, nAbT may be significantly higher in vaccine recipients previously infected with SARS-CoV-2. Monitoring viral evolution is critical to evaluate the impact of novel SARS-CoV-2 variants on vaccine effectiveness, as mutational profiles in the sub-consensus genome could indicate increases in transmissibility and virulence or suggest the development of antiviral resistance.