List your company

Mathematical Reasoning on AIME 24 (# traces, Consistency, Accuracy)

1,206# TracesUniform
38.08341.29644.5947.71Feb 18, 2026

Evaluation results

MethodLinks
UniformBackbone=GPT-20B, Confidence-weighted=true2026.021,2069470.8
UniformBackbone=Qwen3-4B, Confidence-weighted=true2026.029429263.4
UniformBackbone=Qwen3-4B, Confidence-weighted=false2026.028619463
UniformBackbone=GPT-20B, Confidence-weighted=false2026.026659671.9
UniformBackbone=GPT-120B, Confidence-weighted=true2026.025359571.9
UniformBackbone=GPT-120B, Confidence-weighted=false2026.024489773.7
PETS-OfflineBackbone=Qwen3-4B, Confidence-weighted=true2026.023699666
PETS-OfflineBackbone=Qwen3-4B, Confidence-weighted=false2026.0225910063.1
PETS-OfflineBackbone=GPT-20B, Confidence-weighted=true2026.022329772.6
PETS-OfflineBackbone=GPT-120B, Confidence-weighted=true2026.022189973.7
UniformBackbone=Qwen3-30B, Confidence-weighted=false2026.022029869.6
PETS-OfflineBackbone=GPT-20B, Confidence-weighted=false2026.0220010073.3
PETS-OfflineBackbone=GPT-120B, Confidence-weighted=false2026.0218410074.4
UniformBackbone=Qwen3-30B, Confidence-weighted=true2026.021799768.6
PETS-OfflineBackbone=Qwen3-30B, Confidence-weighted=false2026.0213510070
PETS-OfflineBackbone=Qwen3-30B, Confidence-weighted=true2026.0212010069.9
UniformBackbone=Qwen-Long, Confidence-weighted=true2026.021159869.7
UniformBackbone=Qwen-Long, Confidence-weighted=false2026.02969969.9
PETS-OfflineBackbone=Qwen-Long, Confidence-weighted=true2026.02919969.9
PETS-OfflineBackbone=Qwen-Long, Confidence-weighted=false2026.028310070