Starter version, in progress Inspired by the style of Humanity’s Last Exam (HLE), a very hard benchmark for AI. Our questions are original and far easier, written for everyday people. They are not from the real HLE, and we are not affiliated with it.
How do AI models do?
For current AI scores on the real, much harder exam, see the official HLE results and Artificial Analysis. We don’t publish an AI score for this quiz because we haven’t measured one yet.