List your company

Mathematical Reasoning on AMC AIME (test)

67.4AccuracyPIR
41.337648.103854.8761.6362Feb 16, 2026

Evaluation results

MethodLinks
PIRBackbone=DeepSeek-R1-32B2026.0267.42,935.83
Specutive ThinkingBackbone=DeepSeek-R1-32B2026.0264.64,941.73
QwQ-32BBackbone=QwQ-32B2026.02634,600.75
L1-maxBackbone=DeepSeek-R1-32B2026.0260.63,832.18
DeepSeek-R1-32BBackbone=DeepSeek-R1-32B2026.0259.923,661.52
DeepSeek-R1-7BBackbone=DeepSeek-R1-7B2026.0258.333,858.16
DeepSeek-R1-14BBackbone=DeepSeek-R1-14B2026.0257.763,523.09
prompt to conciseBackbone=DeepSeek-R1-32B2026.0254.53,286.95
s1-32BBackbone=DeepSeek-R1-32B2026.0253.965,958.03
DEERBackbone=DeepSeek-R1-32B2026.0251.934,112.46
SpecReasonBackbone=DeepSeek-R1-32B2026.02436,354.78
SEALBackbone=DeepSeek-R1-32B2026.0242.346,354.78