Hacker News Reader

Do you know if OpenAI used the same grading criteria as official judges?

As IMO medalists they would be expected to I'm sure.

But this can be verified because the results are public: