Left-turn gap acceptance at an uncontrolled intersection is dependent on many factors. Most existing studies evaluate gap acceptance in developed countries and less deal with developing countries. Morover, most existing studies also do not differentiate between gap and lag when evaluating gap acceptance. In this paper, a binary logit left-turn gap acceptance model is developed using 1496 field observations in Egypt as one of the developing countries. Gap acceptance behavior was found to be influenced by the type of gap presented to the driver (gap or lag). Size of time interval, driver's time-to-turn, and oncoming driver yielding behavior were found to be the potential influencing factors. Equations for estimating the critical gap and lag were developed and applied for different scenarios. Critical gap and lag were found to be varying over a wide range of. These values were less than those of developed countries which confirms the more risky behavior of drivers in developing countries. The findings from this study can improve operational analysis of left turns at unsignalized intersections by using different critical gaps for different traffic and geometric conditions.