List your company

Mathematical Reasoning on MATH 500 (Accuracy and Latency)

76.2Accuracy (pass@1)Power-SMC
38.55248.32658.167.874Feb 10, 2026

Evaluation results

MethodLinks
Power-SMCModel=Qwen2.5-Math-7B2026.0276.21.44
Scalable Power SamplingModel=Qwen2.5-Math-7B2026.0275.8—
MH power samplingModel=Qwen2.5-Math-7B2026.0274.818.2
Power-SMCModel=Qwen2.5-7B2026.0271.41.64
Scalable Power SamplingModel=Qwen2.5-7B2026.0270.8—
MH power samplingModel=Qwen2.5-7B2026.0270.628.3
Low-temperature decodingModel=Qwen2.5-Math-7B, temperature=τ=1/α2026.02691
Low-temperature decodingModel=Qwen2.5-7B, temperature=τ=1/α2026.0262.81.24
Power-SMCModel=Phi-3.5-mini-instruct2026.0251.63.25
MH power samplingModel=Phi-3.5-mini-instruct2026.0250.816.12
Baseline decodingModel=Qwen2.5-7B2026.0249.81
Baseline decodingModel=Qwen2.5-Math-7B2026.0249.61
Low-temperature decodingModel=Phi-3.5-mini-instruct, temperature=τ=1/α2026.0247.80.91
Baseline decodingModel=Phi-3.5-mini-instruct2026.02401