List your company
ResearchFormerly WizWand

Mathematical Reasoning on AIME 2025 (Pass@1, LEN, TE)

67.3Pass@1Qwen3-8B
9.0624.1839.354.42Jan 13, 2026Jan 17, 2026Jan 22, 2026Jan 26, 2026Jan 31, 2026Feb 4, 2026Feb 9, 2026
Updated 8mo ago

Evaluation results

MethodLinks
Qwen3-8BBackbone=Qwen3-8B2026.0267.318,3423.67
GRPOBackbone=Qwen3-8B2026.0266.613,9814.76
SAGE-GRPOBackbone=Qwen3-8B2026.0266.610,0526.58
GSPOBackbone=Qwen3-8B2026.0266.214,0824.7
SAGE-GSPOBackbone=Qwen3-8B2026.02669,1837.19
GRPOBackbone=DS-7B2026.0238.410,1233.79
SAGE-GRPOBackbone=DS-7B2026.02386,5835.77
DS-7BBackbone=DS-7B2026.0237.112,5402.96
Efficient ReasoningBackbone=DS-7B2026.0236.27,5014.82
GRPO-LEADBackbone=DS-7B2026.0236.17,8424.6
LC-R1Backbone=DS-7B2026.0235.77,4584.79
AdaptThinkBackbone=DS-7B2026.02357,8074.48
GRPOBackbone=DeepScaleR2026.0227.48,1853.35
SAGE-GRPOBackbone=DeepScaleR2026.0227.27,7043.53
SAGE-GSPOBackbone=DS-1.5B2026.0227.17,1673.78
SAGE-GRPOBackbone=DS-1.5B2026.0226.57,4793.54
ThinkPrune-2kBackbone=DeepScaleR2026.02267,4863.47
DeepScaleRBackbone=DeepScaleR2026.0225.49,3102.73
GSPOBackbone=DS-1.5B2026.0225.18,2273.05
GRPOBackbone=DS-1.5B2026.0224.18,2632.92
Efficient ReasoningBackbone=DS-1.5B2026.0222.98,5902.67
AdaptThinkBackbone=DS-1.5B2026.0221.88,1552.67
DS-1.5BBackbone=DS-1.5B2026.0220.911,6691.79
LC-R1Backbone=DS-1.5B2026.0220.96,9423.01
ThinkPrune-2kBackbone=DS-1.5B2026.0219.76,9182.85
Multiplex ThinkingBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0119.7——
Stochastic Soft ThinkingBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0119.1——
Discrete RLBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0117.1——
Discrete CoTBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0116——
Stochastic Soft ThinkingBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0113.2——
Multiplex ThinkingBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0112.8——
Discrete RLBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0112——
Discrete CoTBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0111.3——