Comment by crthpl

GPT-4o is so far behind the frontier; you shouldn't use it as an indicator of what LLMs are capable of.