List your company

Mathematical Reasoning on HMMT25 (mean@8)

13.33Mean Score (8 samples)ACPO
4.87487.06999.26511.4601Jul 3, 2026

Evaluation results

MethodLinks
ACPOModel=DeepSeek-R1-Distill-Qwen-7B2026.0713.33
GTPOModel=DeepSeek-R1-Distill-Qwen-7B2026.0713.19
SAPOModel=DeepSeek-R1-Distill-Qwen-7B2026.0712.46
ACPOModel=Qwen3-8B-Base2026.0711.04
80/20Model=DeepSeek-R1-Distill-Qwen-7B2026.0710.83
SAPOModel=Qwen3-8B-Base2026.0710
ACPOModel=Qwen2.5-Math-7B2026.079.86
DAPOModel=DeepSeek-R1-Distill-Qwen-7B2026.079.61
GTPOModel=Qwen3-8B-Base2026.078.83
GRPOModel=DeepSeek-R1-Distill-Qwen-7B2026.078.75
GTPOModel=Qwen2.5-Math-7B2026.078.7
SAPOModel=Qwen2.5-Math-7B2026.078.64
80/20Model=Qwen3-8B-Base2026.077.91
80/20Model=Qwen2.5-Math-7B2026.077.7
DAPOModel=Qwen3-8B-Base2026.077.12
DAPOModel=Qwen2.5-Math-7B2026.076.67
GRPOModel=Qwen3-8B-Base2026.076.67
GRPOModel=Qwen2.5-Math-7B2026.075.2