List your company

Mathematical Reasoning on AIME24, Math500, Olympiad, AMC, Minerva

15.6AIME24 ScoreDARS-HW-Breadth
11.75212.75113.7514.749Aug 19, 2025

Evaluation results

MethodLinks
DARS-HW-BreadthModel=Qwen2.5-7B-Instruct, Training Data=Math12k2025.0815.676.538.444.734.649.682.3
RLVR baselineModel=Qwen2.5-7B-Instruct, Training Data=Math12k2025.0814.274.837.643.433.448.678.8
Qwen2.5-7B-InstructModel=Qwen2.5-7B-Instruct2025.0811.972.337.142.231.947.280.3