Math Olympiad Solver helps an agent tackle difficult competition math problems from the IMO, Putnam, USAMO, and AIME by combining pure reasoning with rigorous adversarial verification. It uses an interpretation check to catch tricky phrasings, runs parallel solver agents through internal refinement, and strips away thinking traces to prevent agreement bias. A fresh-context adversarial verifier then attacks the proof using specific failure patterns rather than generic checks, outputting calibrated confidence or abstaining when solutions are uncertain. Reach for it when your agent needs to solve complex mathematical proofs, check competition work, or evaluate whether a mathematical proof is actually correct.