For any announcement about performance on the IMO 2025, within 2 months of the IMO 2025.

If the system with the best score is a system that uses both AlphaProof-like AI and o3-like AI, and it is known which fraction of problems were solved by each, I will resolve this market to a fraction relative to each type of system's # of points obtained.

It will be up to my judgement (so I won't take a position), but I will base my positions on known technical details and discussions with people more knowledgeable than me, if the result is ambiguous. May resolve N/A in the worst cases, such as if a company verifiably achieved the best score but doesn't share their method at all.