The equiatomic face-centred cubic (FCC) CoNiCrFeMn alloy, known as the Cantor alloy, is renowned for its high ductility under extreme conditions, such as cryogenic temperatures. Despite this, it suffers from low hardness and yield strength (YS) and includes elements with significant supply concerns. This study introduces novel non-equiatomic CoNiCrFeMn alloys, designed using machine learning (ML)-assisted-high-throughput atomistic simulations to enhance sustainability and mechanical properties such as hardness and YS while maintaining the alloy's single-phase FCC structure. We incorporated two critical sustainability indicators, end-of-life recycling rate (EOL−RR) and lifetime (τ), to guide the alloy design process. Mean-flow stress (MFS), measured from the yield point to 15% strain during tensile tests, was used to assess and predict the mechanical properties. Two alloys, with cobalt contents of 12.5% and 21.5%, were developed and analysed. The Co12.5 alloy showed better sustainability (44% improvement in τ with almost the same mechanical properties), while the Co21.5 alloy exhibited better mechanical properties (20% improvement in MFS with almost the same sustainability) as the equiatomic system. Their stable FCC microstructures were confirmed through CALPHAD modelling and X-ray diffraction (XRD) analysis of vacuum arc melted, as-cast samples. The results highlight the potential of integrating sustainability metrics into high-performance alloy design.