List your company

Mathematical Reasoning on OlympiadBench (Pass@1, Avg)

66.91Pass@1Teacher
13.682827.501441.3255.1386Apr 3, 2026

Evaluation results

MethodLinks
TeacherBackbone=QwQ-32B, Prompting=zero-shot2026.0466.9173.9
MoRSDBackbone=Qwen2.5-Math-1.5B, Prompting=zero-shot2026.0431.7533.07
Gen-SSDBackbone=Qwen2.5-Math-1.5B, Prompting=zero-shot, Chunk size=4K2026.0430.7134.94
Vocabulary Dropout (gen-only)Backbone=Qwen3-8B, Vocabulary Dropout (alpha)=0.75, Phase=gen-only2026.0429.2—
Vocabulary DropoutBackbone=Qwen3-8B, Vocabulary Dropout (alpha)=0.752026.0427.7—
MCC-KDBackbone=Qwen2.5-Math-1.5B, Prompting=zero-shot2026.0427.1530.25
Vocabulary Dropout (train-only)Backbone=Qwen3-8B, Vocabulary Dropout (alpha)=0.75, Phase=train-only2026.0427—
Vocabulary DropoutBackbone=Qwen3-8B, Vocabulary Dropout (alpha)=0.852026.0426.8—
Vocabulary DropoutBackbone=Qwen3-4B, Vocabulary Dropout (alpha)=0.752026.0424.8—
Standard KDBackbone=Qwen2.5-Math-1.5B, Prompting=zero-shot2026.0424.7829.01
Vocabulary DropoutBackbone=Qwen3-4B, Vocabulary Dropout (alpha)=0.852026.0424.7—
Vocabulary Dropout (gen-only)Backbone=Qwen3-4B, Vocabulary Dropout (alpha)=0.75, Phase=gen-only2026.0424.1—
Vocabulary Dropout (train-only)Backbone=Qwen3-4B, Vocabulary Dropout (alpha)=0.75, Phase=train-only2026.0423.6—
BaselineBackbone=Qwen3-4B, Vocabulary Dropout (alpha)=1.02026.0423.2—
BaselineBackbone=Qwen3-8B, Vocabulary Dropout (alpha)=1.02026.0423.1—
Self-DistillationBackbone=Qwen2.5-Math-1.5B, Prompting=zero-shot2026.0422.5524.25
Base (no training)Backbone=Qwen3-4B2026.0420.6—
Base (no training)Backbone=Qwen3-8B2026.0418.1—
Cold StartBackbone=Qwen2.5-Math-1.5B, Prompting=zero-shot2026.0416.1718.3
StudentBackbone=Qwen2.5-Math-1.5B, Prompting=two-shot2026.0415.7315.54