List your company
ResearchFormerly WizWand

Mathematical Reasoning on AIME25 (Accuracy, Average output length)

66.51AccuracyMath-Shepherd-7B
29.621239.198148.77558.3519Mar 1, 2026
Updated 8mo ago

Evaluation results

MethodLinks
Math-Shepherd-7BBackbone=QWEN3-4B, Decoding Strategy=BEST-OF-N, N (Sample Count)=1282026.0366.515,824
OTVBackbone=QWEN3-4B, Decoding Strategy=BEST-OF-N, N (Sample Count)=1282026.0353.915,416
Qwen2.5-PRM-7BBackbone=QWEN3-4B, Decoding Strategy=BEST-OF-N, N (Sample Count)=1282026.0353.333,948
OTVBackbone=DAPO-QWEN-32B, Decoding Strategy=DROP@10, N (Sample Count)=1282026.0350.682,926
Qwen2.5-PRM800K-7BBackbone=QWEN3-4B, Decoding Strategy=STOP@600, N (Sample Count)=1282026.0349.536,409
OTVBackbone=QWEN3-4B, Decoding Strategy=STOP@600, N (Sample Count)=1282026.0349.116,542
OTVBackbone=QWEN3-4B, Decoding Strategy=HALVE@300, N (Sample Count)=1282026.0349.023,170
OTVBackbone=DAPO-QWEN-32B, Decoding Strategy=STOP@600, N (Sample Count)=1282026.0348.232,577
Qwen2.5-PRM800K-7BBackbone=DAPO-QWEN-32B, Decoding Strategy=BEST-OF-N, N (Sample Count)=1282026.0347.814,426
Math-Shepherd-7BBackbone=QWEN3-4B, Decoding Strategy=STOP@600, N (Sample Count)=1282026.0347.456,218
Qwen2.5-PRM-7BBackbone=DAPO-QWEN-32B, Decoding Strategy=BEST-OF-N, N (Sample Count)=1282026.0347.244,481
OTVBackbone=DAPO-QWEN-32B, Decoding Strategy=BEST-OF-N, N (Sample Count)=1282026.0347.084,079
OTVBackbone=DAPO-QWEN-32B, Decoding Strategy=HALVE@300, N (Sample Count)=1282026.0346.982,991
VersaPRM-8BBackbone=QWEN3-4B, Decoding Strategy=STOP@600, N (Sample Count)=1282026.0346.776,113
OTVBackbone=QWEN3-4B, Decoding Strategy=DROP@10, N (Sample Count)=1282026.0346.463,225
Math-Shepherd-7BBackbone=QWEN3-4B, Decoding Strategy=DROP@10, N (Sample Count)=1282026.0346.256,116
Qwen2.5-PRM-7BBackbone=QWEN3-4B, Decoding Strategy=DROP@10, N (Sample Count)=1282026.0345.836,588
Qwen2.5-PRM-7BBackbone=QWEN3-4B, Decoding Strategy=HALVE@300, N (Sample Count)=1282026.0345.736,416
Math-Shepherd-7BBackbone=QWEN3-4B, Decoding Strategy=HALVE@300, N (Sample Count)=1282026.0345.366,118
Qwen2.5-PRM800K-7BBackbone=QWEN3-4B, Decoding Strategy=HALVE@300, N (Sample Count)=1282026.0345.216,619
Qwen2.5-PRM800K-7BBackbone=QWEN3-4B, Decoding Strategy=BEST-OF-N, N (Sample Count)=1282026.0345.16,202
Qwen2.5-PRM-7BBackbone=QWEN3-4B, Decoding Strategy=STOP@600, N (Sample Count)=1282026.0344.956,304
DeepConfBackbone=QWEN3-4B, Decoding Strategy=HALVE@300, N (Sample Count)=1282026.0343.546,318
VersaPRM-8BBackbone=QWEN3-4B, Decoding Strategy=DROP@10, N (Sample Count)=1282026.0343.446,334
VersaPRM-8BBackbone=QWEN3-4B, Decoding Strategy=BEST-OF-N, N (Sample Count)=1282026.0343.286,132
DeepConfBackbone=QWEN3-4B, Decoding Strategy=BEST-OF-N, N (Sample Count)=1282026.0343.079,322
Qwen2.5-PRM800K-7BBackbone=QWEN3-4B, Decoding Strategy=DROP@10, N (Sample Count)=1282026.0342.866,445
Math-Shepherd-7BBackbone=DAPO-QWEN-32B, Decoding Strategy=HALVE@300, N (Sample Count)=1282026.0342.764,919
Math-Shepherd-7BBackbone=DAPO-QWEN-32B, Decoding Strategy=BEST-OF-N, N (Sample Count)=1282026.0342.44,570
Qwen2.5-PRM800K-7BBackbone=DAPO-QWEN-32B, Decoding Strategy=DROP@10, N (Sample Count)=1282026.0341.24,689
DeepConfBackbone=QWEN3-4B, Decoding Strategy=STOP@600, N (Sample Count)=1282026.0340.897,084
Qwen2.5-PRM800K-7BBackbone=DAPO-QWEN-32B, Decoding Strategy=STOP@600, N (Sample Count)=1282026.0340.894,523
DeepConfBackbone=QWEN3-4B, Decoding Strategy=DROP@10, N (Sample Count)=1282026.0340.787,260
Qwen2.5-PRM800K-7BBackbone=DAPO-QWEN-32B, Decoding Strategy=HALVE@300, N (Sample Count)=1282026.03404,528
VersaPRM-8BBackbone=DAPO-QWEN-32B, Decoding Strategy=DROP@10, N (Sample Count)=1282026.0339.484,796
Math-Shepherd-7BBackbone=DAPO-QWEN-32B, Decoding Strategy=DROP@10, N (Sample Count)=1282026.0339.224,475
DeepConfBackbone=DAPO-QWEN-32B, Decoding Strategy=BEST-OF-N, N (Sample Count)=1282026.0338.913,957
DeepConfBackbone=DAPO-QWEN-32B, Decoding Strategy=STOP@600, N (Sample Count)=1282026.0337.764,353
VersaPRM-8BBackbone=DAPO-QWEN-32B, Decoding Strategy=STOP@600, N (Sample Count)=1282026.0337.665,005
VersaPRM-8BBackbone=QWEN3-4B, Decoding Strategy=HALVE@300, N (Sample Count)=1282026.0337.246,438
DeepConfBackbone=DAPO-QWEN-32B, Decoding Strategy=DROP@10, N (Sample Count)=1282026.0337.084,398
DeepConfBackbone=DAPO-QWEN-32B, Decoding Strategy=HALVE@300, N (Sample Count)=1282026.0336.774,449
VersaPRM-8BBackbone=DAPO-QWEN-32B, Decoding Strategy=HALVE@300, N (Sample Count)=1282026.0336.615,046
Math-Shepherd-7BBackbone=DAPO-QWEN-32B, Decoding Strategy=STOP@600, N (Sample Count)=1282026.0336.465,063
Qwen2.5-PRM-7BBackbone=DAPO-QWEN-32B, Decoding Strategy=HALVE@300, N (Sample Count)=1282026.0336.094,967
Qwen2.5-PRM-7BBackbone=DAPO-QWEN-32B, Decoding Strategy=STOP@600, N (Sample Count)=1282026.0335.315,068
Qwen2.5-PRM-7BBackbone=DAPO-QWEN-32B, Decoding Strategy=DROP@10, N (Sample Count)=1282026.0333.655,351
VersaPRM-8BBackbone=DAPO-QWEN-32B, Decoding Strategy=BEST-OF-N, N (Sample Count)=1282026.0331.044,447