List your company
ResearchFormerly WizWand

Math Reasoning on OlympiadBench (Pass@1 accuracy)

48.2Pass@1 Accuracys1.1-7B
22.228.9535.742.45Dec 16, 2025Dec 20, 2025Dec 25, 2025Dec 30, 2025Jan 3, 2026Jan 8, 2026Jan 13, 2026
Updated 8mo ago

Evaluation results

MethodLinks
s1.1-7BBackbone=Qwen2.5-7B, Pure SFT?=false2025.1248.2
Multiplex ThinkingBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0141.7
Stochastic Soft ThinkingBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0140.6
Qwen InstructBackbone=Qwen2.5-7B, Pure SFT?=false2025.1238.7
Discrete RLBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0138
Qwen InstructBackbone=Qwen2.5-Math-1.5B, Pure SFT?=false2025.1237.9
Discrete CoTBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0135.6
LadderBackbone=Qwen2.5-7B, Pure SFT?=true2025.1234.3
QLoRABackbone=Qwen2.5-Math-1.5B, Pure SFT?=true2025.1233.7
LadderBackbone=Qwen2.5-Math-1.5B, Pure SFT?=true2025.1232.7
QLoRABackbone=Qwen2.5-7B, Pure SFT?=true2025.1232
Multiplex ThinkingBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0131.3
Discrete RLBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0131.2
Stochastic Soft ThinkingBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0130.6
Discrete CoTBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0130.5
Qwen BaseBackbone=Qwen2.5-7B, Pure SFT?=true2025.1230.2
Full SFTBackbone=Qwen2.5-7B, Pure SFT?=false2025.1227.6
Qwen BaseBackbone=Qwen2.5-Math-1.5B, Pure SFT?=true2025.1223.8
Full SFTBackbone=Qwen2.5-Math-1.5B, Pure SFT?=true2025.1223.2