The benchmark is not everything, the LLMs have their “personality” and GPT is annoying AF.

Also, I don’t about others, but I personally strongly dislike OpenAI’s leadership’s hypocrisy. I find them losing the race highly satisfying.