List your company

Mathematical Reasoning on OlympiadBench (Avg@K)

44.25Avg Success Rate @8IntroLLM
11.427619.948828.4736.9912Dec 4, 2025Dec 15, 2025Dec 27, 2025Jan 8, 2026Jan 20, 2026Feb 1, 2026Feb 13, 2026

Evaluation results

MethodLinks
IntroLLMBackbone=Qwen3-4B-Base2026.0244.25——
GRPOBackbone=Qwen3-4B-Base, Temperature (τ)=1.22026.0244.12——
EADBackbone=Qwen3-4B-Base, Schedule=1.2→0.12026.0243.69——
SENTBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.1242.8542.9942.95
w/ MaskBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.1242.6542.8942.88
TAMPOBackbone=Qwen3-4B-Base, Range=0.6-1.52026.0242.43——
GRPOBackbone=Qwen3-4B-Base, Temperature (τ)=1.02026.0242.36——
w/ ClipBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.1241.4441.3641.25
GRPOBackbone=Qwen3-4B-Base, Temperature (τ)=0.62026.0241.06——
w/ EnBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.1240.4840.1140.39
w/ CovBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.1239.7240.2940.92
w/ High EnBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.1239.5939.739.48
w/ AdvBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.1237.0436.2836.24
Base modelBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.1235.6536.5936.51
IntroLLMBackbone=Qwen3-1.7B-Base2026.0229.95——
GRPOBackbone=Qwen3-1.7B-Base, Temperature (τ)=1.22026.0229.93——
EADBackbone=Qwen3-1.7B-Base, Schedule=1.2→0.12026.0229.64——
GRPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.1229.4829.4329.6
GRPOBackbone=Qwen3-1.7B-Base, Temperature (τ)=1.02026.0228.78——
GRPOBackbone=Qwen3-1.7B-Base, Temperature (τ)=0.62026.0228.23——
TAMPOBackbone=Qwen3-1.7B-Base, Range=0.6-1.52026.0227.41——
BaseBackbone=Qwen3-4B-Base2026.0224.3——
BaseBackbone=Qwen3-1.7B-Base2026.0212.69——
OpenPangu-embedded-1B-v1.1Base Model=DeepSeek-R1-Distill-Qwen-1.5B2025.12—36.3—
SENTBase Model=DeepSeek-R1-Distill-Qwen-1.5B2025.12—42.99—