List your company

Scientific Q&A on GPQA Diamond (test)

65.22AccuracyPIR
21.103232.556644.0155.4634Feb 16, 2026

Evaluation results

MethodLinks
PIRBackbone=DeepSeek-R1-32B2026.0265.223,024.81
s1-32BBackbone=DeepSeek-R1-32B2026.0260.723,402.53
SEALBackbone=DeepSeek-R1-32B2026.0256.133,468.96
DEERBackbone=DeepSeek-R1-32B2026.0254.23,402.51
Specutive ThinkingBackbone=DeepSeek-R1-32B2026.0251.375,031.91
SpecReasonBackbone=DeepSeek-R1-32B2026.0244.434,885.98
QwQ-32BBackbone=QwQ-32B2026.0237.425,585.19
DeepSeek-R1-32BBackbone=DeepSeek-R1-32B2026.0236.35,028.59
prompt to conciseBackbone=DeepSeek-R1-32B2026.0231.984,120.47
L1-maxBackbone=DeepSeek-R1-32B2026.0225.813,769.06
DeepSeek-R1-14BBackbone=DeepSeek-R1-14B2026.0224.664,411.01
DeepSeek-R1-7BBackbone=DeepSeek-R1-7B2026.0222.85,020.86