BMAT results (October 2018).

The table shows the results obtained by ChatGPT-3.5 (left) and ChatGPT-4 (right) in the October 2018 BMAT test. For Sections 1 and 2 of the test, we report the number of questions administered, the raw scores (correct answers/questions administered) in each subject area and in the whole section, the normalized raw score, and the corresponding BMAT score followed by the relative percentile. At the bottom of each section, the UCL applicants, interviewed, and offer holders average scores are reported. The bottom row shows the ratio of the total number of correct answers to the total number of questions administered and the corresponding percentage.