How sleep impacts the accuracy of identifications that eyewitnesses make from lineups is unknown. For a comprehensive understanding of eyewitness performance, two types of eyewitness ID accuracy are considered: discriminability (the ability to distinguish innocent from guilty suspects) and reliability (the probability that the identified suspect was the offender). The well-known role sleep plays in memory consolidation should apply to an eyewitness’s ability to discriminate, but not necessarily their reliability. That is what we investigated in a large-scale forensically-relevant experiment. We compared discriminability and reliability from sleep (sleep occurs between witnessing a crime and lineup test) and wake (remains awake between crime and lineup) conditions. Furthermore, theorists have long been using signal-detection models to understand recognition memory, but its use is new to the field of eyewitness ID research. Thus, we compared signal-detection models with different decision rules. Our findings shed light on the impact sleep has on eyewitness IDs.