List your company
ResearchFormerly WizWand

Math Reasoning on HMMT25 (Accuracy)

34.9Accuracy (HMMT25)Uni-OPD
11.39617.49823.629.702Sep 25, 2025Nov 1, 2025Dec 8, 2025Jan 14, 2026Feb 20, 2026Mar 29, 2026May 5, 2026
Updated 6mo ago

Evaluation results

MethodLinks
Uni-OPDDistillation Scenario=Multi-Teacher Distillation, Model Scale=4B2026.0534.9
Uni-OPDDistillation Scenario=Single-Teacher Distillation, Model Scale=4B2026.0534.8
ExOPDDistillation Scenario=Multi-Teacher Distillation, Model Scale=4B2026.0534.4
ExOPDDistillation Scenario=Single-Teacher Distillation, Model Scale=4B2026.0533.9
OPDDistillation Scenario=Multi-Teacher Distillation, Model Scale=4B2026.0533.4
TeacherTraining Strategy=RL2026.0532.5
ExPODistillation Scenario=Single-Teacher Distillation, Model Scale=4B2026.0532.4
ExPODistillation Scenario=Multi-Teacher Distillation, Model Scale=4B2026.0531.7
SFTDistillation Scenario=Multi-Teacher Distillation, Model Scale=4B2026.0530.7
CE-GPPOBase Model=DS-R1-Distill-Qwen-7B, Training Algorithm=CE-GPPO, Beta1 Hyperparameter=0.75, Beta2 Hyperparameter=12025.0930.5
OPDDistillation Scenario=Single-Teacher Distillation, Model Scale=4B2026.0530.2
CE-GPPOBase Model=DS-R1-Distill-Qwen-7B, Training Algorithm=CE-GPPO, Beta1 Hyperparameter=0.5, Beta2 Hyperparameter=12025.0927.3
DS-R1-Distill-Qwen-7BBase Model=DS-R1-Distill-Qwen-7B2025.0926.2
DAPOBase Model=DS-R1-Distill-Qwen-7B, Training Algorithm=DAPO2025.0925.6
GRPOBase Model=DS-R1-Distill-Qwen-7B, Training Algorithm=GRPO2025.0924.5
CE-GPPOBase Model=DS-R1-Distill-Qwen-1.5B, Training Algorithm=CE-GPPO, Beta1 Hyperparameter=0.5, Beta2 Hyperparameter=12025.0921.6
CE-GPPOBase Model=DS-R1-Distill-Qwen-1.5B, Training Algorithm=CE-GPPO, Beta1 Hyperparameter=0.75, Beta2 Hyperparameter=12025.0919.3
DAPOBase Model=DS-R1-Distill-Qwen-1.5B, Training Algorithm=DAPO2025.0919.2
GRPOBase Model=DS-R1-Distill-Qwen-1.5B, Training Algorithm=GRPO2025.0916.6
DS-R1-Distill-Qwen-1.5BBase Model=DS-R1-Distill-Qwen-1.5B2025.0913.1
StudentModel Scale=4B2026.0512.3