List your company

Mathematical Reasoning on MATH 500 (Accuracy, Gen. Cost)

94.1AccuracyZIP-RC sampling
81.41284.7068891.294Dec 1, 2025

Evaluation results

MethodLinks
ZIP-RC samplingModel=Qwen3-1.7B2025.1294.11.43
Weighted BoN Self-evalModel=Qwen3-1.7B2025.1293.61.4
ZIP-RC samplingModel=LFM2-1.2B2025.1293.41.35
Majority VotingModel=Qwen3-1.7B2025.12931.4
Weighted BoN ext. RMModel=Qwen3-1.7B2025.1292.61.43
Weighted BoN Self-evalModel=LFM2-1.2B2025.1292.61.6
ZIP-RC reward pruneModel=LFM2-1.2B2025.1292.51.49
Majority VotingModel=LFM2-1.2B2025.1291.81.6
MV length-pruneModel=LFM2-1.2B2025.1291.61.7
Weighted BoN ext. RMModel=LFM2-1.2B2025.1291.11.53
ZIP-RC reward pruneModel=Qwen3-1.7B2025.1290.31.33
ZIP-RC samplingModel=LFM2-350M2025.1286.11.49
MV length-pruneModel=Qwen3-1.7B2025.1284.71.46
Weighted BoN Self-evalModel=LFM2-350M2025.1284.41.7
MV length-pruneModel=LFM2-350M2025.1283.61.66
ZIP-RC reward pruneModel=LFM2-350M2025.1283.21.27
Majority VotingModel=LFM2-350M2025.1282.71.7
Weighted BoN ext. RMModel=LFM2-350M2025.1281.91.59