Global Navigation Satellite System (GNSS) timing is a main service function. Each GNSS has its own time performance specification. However, a uniform timing performance assessment methodology and its outcomes do not exist. Firstly, the timing performance specifications of each GNSS are analyzed. Then, time transfer accuracy is considered as the key GNSS timing performance indicator. Secondly, an assessment method for the Coordinated Universal Time (UTC) published by the Bureau International des Poids et Mesures (BIPM) and UTC kept by the National Time Service Center of China (UTC(NTSC)) is proposed. Thirdly, the uncertainty budget of the assessment method is given. The timing performances of BDS, GPS, GLONASS, and Galileo are assessed and compared. The results show that the time transfer accuracy of BDS, GPS, GLONASS, and Galileo was 13.8 ns, 4.5 ns, 16.8 ns, and 4.2 ns, respectively, in 2021, meeting their performance requirements specified by GNSS (30 ns or 40 ns). Meanwhile, the assessment results of GPS and Galileo are much better than requirements, while the assessment results of BDS and GLONASS show fixed time offset and can still be improved further. If the local reference time of GNSS users can be connected with UTC, this assessment method can be used.

The receiver delay has a significant impact on global navigation satellite system (GNSS) time measurement. This article comprehensively analyzes the difficulty, composition, principle, and calculation of GNSS receiver delay. A universal method, based on clock-steering characterization, is proposed to absolutely calibrate all types of receivers. We use a hardware simulator to design several experiments to test the performance of GNSS receiver delay for different receiver types, radio frequency (RF) signals, operation status and time-to-phase (TtP). At first, through the receivers of Novatel and Septentrio, the channel delay of Septentrio is 2 ns far lower than 65 ns for Novatel, and for the inter-frequency bias of GLONASS L1, Septentrio tends to increase within 10 ns compared with decreasing of Novatel within 5 ns. Secondly, a representative receiver of UniNav-BDS (BeiDou) is chosen to test the influence of Ttp which may be ignored by users. Under continuous operation, the receiver delay shows a monotone reduction of 10 ns as TtP increased by 10 ns. However, under on-off operation, the receiver delay represents periodic variation. Through a zero-baseline comparison, we verifies the relation between receiver delay and TtP. At last, the article analyzes instrument errors and measurement errors in the experiment, and the combined uncertainty of absolute calibration is calculated with 1.36 ns.