List your company

Mathematical Reasoning on AIME 2025 (Mean@32)

16.51Mean Score (32 Samples)Qwen2.5-Math-7B + Resample w/ LOPE (w/ Training Signal Shaping)
1.41965.33739.25513.1727May 7, 2026

Evaluation results

MethodLinks
Qwen2.5-Math-7B + Resample w/ LOPE (w/ Training Signal Shaping)Model=Qwen2.5-Math-7B, Optimization Method=GRPO + Resample, Prompt Strategy=LOPE, Training Signal Shaping=true2026.0516.51
Qwen3-4B-Base + Resample w/ LOPE (w/ Training Signal Shaping)Model=Qwen3-4B-Base, Optimization Method=GRPO + Resample, Prompt Strategy=LOPE, Training Signal Shaping=true2026.0516.27
Qwen3-4B-Base + Resample w/ LOPE (w/o Training Signal Shaping)Model=Qwen3-4B-Base, Optimization Method=GRPO + Resample, Prompt Strategy=LOPE, Training Signal Shaping=false2026.0513.85
Qwen3-4B-Base + GRPOModel=Qwen3-4B-Base, Optimization Method=GRPO2026.0513.12
Qwen3-4B-Base + Resample w/ Naive PromptModel=Qwen3-4B-Base, Optimization Method=GRPO + Resample, Prompt Strategy=Naive Prompt2026.0511.67
Qwen2.5-Math-7B + Resample w/ LOPE (w/o Training Signal Shaping)Model=Qwen2.5-Math-7B, Optimization Method=GRPO + Resample, Prompt Strategy=LOPE, Training Signal Shaping=false2026.0510.52
Qwen2.5-Math-7B + GRPOModel=Qwen2.5-Math-7B, Optimization Method=GRPO2026.059.9
Qwen2.5-Math-7B + Resample w/ Naive PromptModel=Qwen2.5-Math-7B, Optimization Method=GRPO + Resample, Prompt Strategy=Naive Prompt2026.059.17
Qwen2.5-Math-7BModel=Qwen2.5-Math-7B, Optimization Method=Base2026.057.9
Qwen3-4B-BaseModel=Qwen3-4B-Base, Optimization Method=Base2026.057.24
Qwen3-1.7B-Base + Resample w/ LOPE (w/o Training Signal Shaping)Model=Qwen3-1.7B-Base, Optimization Method=GRPO + Resample, Prompt Strategy=LOPE, Training Signal Shaping=false2026.055.83
Qwen3-1.7B-Base + Resample w/ LOPE (w/ Training Signal Shaping)Model=Qwen3-1.7B-Base, Optimization Method=GRPO + Resample, Prompt Strategy=LOPE, Training Signal Shaping=true2026.055.73
Qwen3-1.7B-Base + Resample w/ Naive PromptModel=Qwen3-1.7B-Base, Optimization Method=GRPO + Resample, Prompt Strategy=Naive Prompt2026.054.58
Qwen3-1.7B-Base + GRPOModel=Qwen3-1.7B-Base, Optimization Method=GRPO2026.054.47
Qwen3-1.7B-BaseModel=Qwen3-1.7B-Base, Optimization Method=Base2026.052