Large language models are better than humans at answering chemistry questions

AI models outperform human chemists in every topic area. But are they really better chemists?