ResearchBenchmarksMathematical Reasoning on OlympiadBench (pass@8)Follow83.82Pass@8 AccuracyOPHSD74.647277.028679.4181.7914May 9, 2026Evaluation ResultsMethodMethodLinksPass@8 AccuracyOPHSDharness=noneharness=none2026.0583.82OPSD2026.0580.88OPHSDharness=plan-solveharness=plan-solve2026.0579.85GRPO2026.0577.94CRISP2026.0575