OpenAI claiming gold medal standard at IMO 2025 (github.com)

According to the 6/N from this series, they are claiming full marks for problems 1 -- 5

https://x.com/alexwei_/status/1946477742855532918

I posted about one of the twitter threads at https://news.ycombinator.com/item?id=44613840

The proof superficially look super interesting. Especially bc it's not in style of usual LLM babble fillers. It's like almost exactly opposite, very efficient use of words and eliminating parts of grammar not important. Reminds me of how people write down proofs in drafts/how we communicate proofs with peers before writing final versions.

This is incredible. We know these questions are not in the training data. How can you still say that LLMs aren't reasoning.