Autonomous vehicles are rapidly evolving, but public acceptance remains challenging. Research on mixed traffic flow, where human-driven and self-driving cars coexist, is crucial. Most studies focus on the impact of a single variable on route selection, overlooking the combined influence of multiple factors and varying driver cognition. This study delves into traffic efficiency, safety, and the overall traffic environment to scrutinize the network equilibrium of route selections in a mixed traffic environment. By incorporating the heterogeneity of comprehensive cost cognition into a route choice model, the analysis of Hong Kong’s road network yields several pivotal insights. Firstly, autonomous vehicles exhibit considerable variations in route preferences under diverse scenarios for the same origin-destination pair, highlighting their sensitivity to cost performance. Secondly, the comprehensive travel costs in a mixed traffic environment are significantly higher compared to those in single traffic flow scenarios, revealing a greater heterogeneity in cost cognition. Thirdly, Mixed traffic flow assignment is more balanced. This study provides insights for path selection in mixed traffic, offering valuable guidance for vehicle-road collaboration and promoting green travel.