In this article, we are focusing on the manufacturing techniques of Greek and Roman copper alloy coins (fifth century BC—third century AD). Thanks to the metallographic analysis of the objects remaining today, it is possible to trace their metallurgical history and to reconstruct the different stages of the coin’s manufacturing process in antiquity. The 135 coins studied by various methods (X-ray fluorescence, scanning electron microscopy, Vickers hardness) show a large range of elemental compositions. The most represented alloy in our samples is a ternary alloy of copper, lead and tin, but other alloys were also used (Cu–Pb, Cu–Sn, Cu–Sn–Pb, Cu–Ag, Cu–Zn, and some ‘pure’ copper coins). Coins also show diversity in terms of microstructures, representing either a simple process (casting followed by striking) or a more complex process with a recrystallized microstructure (67% of our sample). The grain size of these recrystallized microstructures provides the metallurgical state of the material. Furthermore, Vickers hardness measurements taken on coins provided us information on the impact of the minting process on the coin but also, depending on the alloys, provides indications on the state of solidification and shaping of the material. The average hardness of cast state alloys is 97.5 HV compared to 119.3 HV for recrystallized samples. Some recrystallized samples with a hardness lower than expected suggest that annealing has probably been carried out before striking. The resulting alloy is easier to form and strike. Sessions of experimental archaeology allowed the production of samples representative of ancient conditions and the comparison of these samples with actual archaeological samples. Hardness measurements on these reproduced samples do not exceed 152 HV, with minimal recrystallization, whereas some archaeological samples (15%) may reach values between 150 and 214 HV. This hardness variation indicates higher deformations due to several cycles of hammering and annealing on archaeological samples that were not previously observed.< Réduire