Accurate detection of incident hepatitis C virus (HCV) infection is required to target and evaluate public health interventions, but acute infection is largely asymptomatic and difficult to detect using traditional methods. Our aim was to evaluate a previously developed HCV avidity assay to distinguish acute from chronic HCV infection. Plasma samples collected from recent seroconversion subjects in two large Australian cohorts were tested using the avidity assay, and the avidity index (AI) was calculated. Demographic and clinical characteristics of patients with low/high AI were compared via logistic regression. Sensitivity and specificity of the assay for recent infection and the mean duration of recent infection (MDRI) were estimated stratified by HCV genotype. Avidity was assessed in 567 samples (from 215 participants), including 304 with viraemia (defined as ≥250 IU/mL). An inverse relationship between AI and infection duration was found in viraemic samples only. The adjusted odds of a low AI (<30%) decreased with infection duration (odds ratio [OR] per week of 0.93; 95% CI:0.89‐0.97), and were lower for G1 compared with G3 samples (OR = 0.14; 95% CI:0.05‐0.39). Defining recent infection as <26 weeks, sensitivity (at AI cut‐off of 20%) was estimated at 48% (95% CI:39‐56%), 36% (95% CI:20‐52%), and 65% (95% CI:54‐75%) and MDRI was 116, 83, and 152 days for all genotypes, G1, and G3, respectively. Specificity (≥52 weeks infection duration, all genotypes) was 96% (95% CI:90‐98%). HCV avidity testing has utility for detecting recent HCV infection in patients, and for assessing progress in reaching incidence targets for eliminating transmission, but variation in assay performance across genotype should be recognized.