Yes I'm aware of it. I meant it more in absolute terms as a reference (60 is 2 times more than 30 no? ;) ) to make the point that the AMC 12 scores are way better than the AMC 10 scores. Nevertheless the bigger point is that there seems to be some anomaly in the test scores. Maybe some data contamination or some bug in their automated test suite. And on twitter quite a few folks also mentioned this, including a former OpenAI engineer[0] who worked on automated theorem proving. I'm pretty sure this will be looked into further in the coming weeks.

[0] https://twitter.com/spolu/status/1635903343397576705