Selling virtual items is one of the major revenue sources for online game operators. Therefore, the design of game currency systems is important. This research shows that the game currency exchange rate (e.g., $1 = 1,000 game currency is a higher rate than $1 = 10 game currency) affects price perception of virtual items. When the rate is not prominent in players’ minds (e.g., the rate is not mentioned or players have sufficient game currency in their account), players consider a virtual item as more expensive if the rate is higher. In contrast, when the rate is prominent in players’ minds (e.g., the rate is mentioned and no additional information indicates that players already have sufficient game currency, and thus it is irrelevant), players consider a virtual item as cheaper if the rate is higher. In the latter case, higher rate leads to stronger purchase intention of nonsocial, but not social, virtual items. Our findings show currency system design can affect players’ willingness to spend on virtual items thus affecting the revenue of game operators. Moreover, reminders of the actual monetary value of virtual items might help to reduce players’ overspending driven by the biasing effect of the game currency exchange rate. Selling virtual items (e.g., clothing, weapons, or virtual furniture) serves as one of the major revenue sources for online game operators. Therefore, exploring what specific factors and design attributes may affect the sales of virtual items has become an important issue. This research draws upon prior studies on currency numerosity, contextualizing them to online game contexts to develop several theory-based predictions about how online game currency numerosity affects players’ perceived expensiveness and consequently their purchase intention of virtual items. To test these predictions, we conducted six experiments that investigate numerosity effects in massive multiplayer online games (MMOs) where players can exchange real money for game currency based on an exchange rate in order to purchase virtual items. Results indicate the following: (1) when the game currency exchange rate is not salient (e.g., game currency exchange rate is not mentioned or players have sufficient game currency in their account), players consider a virtual item as more expensive if the price is high (versus low) in numerosity; but (2) when the game currency exchange rate is salient (e.g., game currency exchange rate is mentioned and no additional information indicates that it is irrelevant), players consider a virtual item as cheaper if the price is high (versus low) in numerosity; in addition, (3) in the latter case, numerosity increases players’ purchase intention of nonsocial, but not social, virtual items through the mediating effect of perceived expensiveness. Our notable findings contribute to the existing literature on numerosity and provide guidance for the design of virtual currency systems in MMOs.