List your company

Mathematical Reasoning on Math Category Average (test)

51.5AccuracyRLVR
12.281622.463332.64542.8267Mar 6, 2026

Evaluation results

MethodLinks
RLVRModel Size=7B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0351.50.630.460.45
RLVR+QuestionProbModel Size=7B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0350.230.320.470.47
RLVR+Assessor ModelModel Size=7B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0350.230.830.240.19
RLVR+ProbeModel Size=7B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0350.230.840.120.1
CoCAModel Size=7B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0347.920.710.10.09
RLVR+QuestionProbModel Size=3B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0346.530.420.440.45
RLVR+Assessor ModelModel Size=3B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0346.530.830.160.14
RLVR+ProbeModel Size=3B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0346.530.840.110.11
CoCAModel Size=3B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0344.850.880.090.09
Instruct ModelModel Size=7B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0343.730.610.520.5
RLVRModel Size=3B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0342.860.340.480.47
RLVR+QuestionProbModel Size=1.5B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0339.40.440.360.37
RLVR+Assessor ModelModel Size=1.5B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0339.40.720.160.14
RLVR+ProbeModel Size=1.5B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0339.40.620.10.12
Instruct ModelModel Size=3B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0337.540.540.540.52
CoCAModel Size=1.5B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0336.850.820.090.1
RLVRModel Size=1.5B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0330.890.520.620.61
Instruct ModelModel Size=1.5B, Backbone=Qwen2.5-Instruct, Temperature=1.0, Evaluation=pass@12026.0313.790.520.770.73