Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Step-level reasoning evaluation on PRMBench step-level 2025
Loading...
47.13
FPR
PRISM
46.2416
52.2383
58.235
64.2317
Jun 8, 2026
FPR
FNR
Precision
TPR
TNR
PRMScore
FPR Reduction
Updated 1mo ago
Evaluation Results
Method
Method
Links
FPR
FNR
Precision
TPR
TNR
PRMScore
FPR Reduction
PRISM
Loss function=Pairwise...
2026.06
47.13
12.69
91.93
87.3
52.86
68
32.22
Qwen-PRM-7B + Pairwise Loss (12k data) + CL
Loss function=Pairwise...
2026.06
60.58
7.2
90.4
92.79
39.42
67.3
12.63
Qwen-PRM-7B + Pairwise Loss (12k data)
Loss function=Pairwise...
2026.06
61.13
7.1
88.96
92.41
38.86
67
11.84
Qwen-PRM-7B + Pointwise Loss + Data Aug. (132k)
Loss function=Pointwis...
2026.06
67
6
89.6
94
33.5
65.6
3.37
Qwen-PRM-7B (Baseline)
Backbone=Qwen-PRM-7B
2026.06
69.34
4.27
89.4
95.36
30.66
65.5
0
Feedback
Search any
task
Search any
task