Receiver code biases (RCBs) are known to be time delays within the receiver caused by their hardware imperfections. To better understand the characteristics of RCBs, the un-combined (UC) and ionosphere-free (IF) precise point positioning functional models are adapted and re-parameterized to estimate the variation of RCBs as a time-variant parameter. In this study, we analytically studied the temporal variations of RCBs; although there exists a benchmark difference between the UC and IF models, their estimates are in accordance with each other. Additionally, this contribution assesses the inter-day stability of RCBs with weekly observations from 165 globally distributed international global navigation satellite system service stations equipped the receivers of three mainly types. The inter-day stability results of RCB revealed that the RCBs of POL2 and OUS2 have better stability over consecutive 7 d and the single differenced (SD) RCBs can reach 0.2 m in the best case. The results show that 74.83% of the stations are equipped with Trimble receivers under the condition that the mean SD RCB values are between −0.5 and 0.5 m, while 85.57% of the stations are equipped with Septentrio receivers and the stations equipped with Javad can reach 84.35% under this condition. The RCB estimates are also relatively stable for the case in which the receiver hardware device stays unchanged. The relationship between RCBs, receiver type, and antenna configuration is found using six groups of receivers. A strong correlation exists between RCBs, receiver type, and antenna configuration, which is more obvious among Septentrio receivers. The results show that the Pearson correlation coefficients were all higher than 0.9, and the standard deviation of between-receiver RCBs was smaller than 0.327 m when equipped with Septentrio receivers. We concluded that there is a strong relationship between the receiver-related pseudorange biases and the receiver and antenna setup.