List your company

Mathematical Reasoning on MATH 500 (Pass@1, LEN, TE)

95Pass@1SAGE-GRPO
65.15272.90180.6588.399Jan 13, 2026Jan 17, 2026Jan 22, 2026Jan 26, 2026Jan 31, 2026Feb 4, 2026Feb 9, 2026

Evaluation results

MethodLinks
SAGE-GRPOBackbone=Qwen3-8B2026.02953,01531.5
GSPOBackbone=Qwen3-8B2026.0294.64,34222.2
Qwen3-8BBackbone=Qwen3-8B2026.0294.45,64016.7
SAGE-GSPOBackbone=Qwen3-8B2026.0294.42,75334.3
GRPOBackbone=Qwen3-8B2026.0293.64,47020.9
SAGE-GRPOBackbone=DS-7B2026.02932,14143.4
GRPOBackbone=DS-7B2026.02923,21928.5
DS-7BBackbone=DS-7B2026.0291.63,87123.7
Efficient ReasoningBackbone=DS-7B2026.0289.82,40837.3
GRPO-LEADBackbone=DS-7B2026.0289.52,75232.5
AdaptThinkBackbone=DS-7B2026.0288.92,19940.4
SAGE-GRPOBackbone=DeepScaleR2026.0288.83,11728.4
GRPOBackbone=DeepScaleR2026.0287.63,48225.2
LC-R1Backbone=DS-7B2026.0287.32,07642.1
DeepScaleRBackbone=DeepScaleR2026.02863,80522.6
SAGE-GSPOBackbone=DS-1.5B2026.0285.22,92129.2
SAGE-GRPOBackbone=DS-1.5B2026.0284.82,91529.1
GRPOBackbone=DS-1.5B2026.0283.63,90721.4
GSPOBackbone=DS-1.5B2026.0283.43,89825.3
DS-1.5BBackbone=DS-1.5B2026.0283.24,88217
ThinkPrune-2kBackbone=DeepScaleR2026.0282.52,94628
Efficient ReasoningBackbone=DS-1.5B2026.02822,82129.1
ThinkPrune-2kBackbone=DS-1.5B2026.0281.72,82628.9
LC-R1Backbone=DS-1.5B2026.0280.42,97327
AdaptThinkBackbone=DS-1.5B2026.0280.42,56331.4
Multiplex ThinkingBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0178——
Stochastic Soft ThinkingBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0176.5——
Discrete RLBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0174.1——
Discrete CoTBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0171.6——
Multiplex ThinkingBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0167.5——
Stochastic Soft ThinkingBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0166.8——
Discrete RLBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0166.7——
Discrete CoTBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0166.3——