Google's answers were judged by IMO. OpenAI's were judged by themselves internally. Whether it matters is up to the reader.