Recent studies have revisited the claim that OpenAI's GPT-4 scored in the 90th percentile on the bar exam, revealing it actually scored in the 69th percentile overall and 48th percentile among passing students. Despite this, GPT-4 represents a significant advancement over GPT-3.5, with applications like CoCounsel from Casetext combining GPT-4 with legal data for reliable results. Various teams have trained custom legal models with OpenAI, showing high preference and accuracy rates compared to other models.