Florian Othnin

ALT The image you provided is an infographic titled "How Smart is ChatGPT?" It compares the performance of OpenAI's language models, ChatGPT 3.5 and ChatGPT 4.0, across various professional and academic exams. The exams are shown along a percentile ranking axis ranging from the 10th to the 99th percentile. The infographic details the percentile ranks for both models in exams such as the GRE (Verbal and Quantitative), SAT (Math and Evidence-based Reading & Writing), LSAT, and several Advanced Placement (AP) exams including Biology, Statistics, Psychology, Chemistry, Physics, Calculus, English Language, and English Literature. It also mentions GPT's performance in a programming context (Codeforces Rating). Key findings: - ChatGPT 4.0 shows improvements over ChatGPT 3.5 in most areas. - The model scored highly (above 80th percentile) in areas like Math SAT, Biology AP, and GRE Verbal. - However, it struggled in some subjects, notably scoring below the 30th percentile in AP English Literatur