List your company

Question Answering on SuperGPQA (test)

40.2AccuracySFT
9.925617.785325.64533.5047Jan 5, 2026

Evaluation results

MethodLinks
SFTBackbone=Qwen32026.0140.2
5-shotBackbone=Qwen32026.0135.56
kNN-MoEBackbone=Qwen32026.0135.15
Zero-shotBackbone=Qwen32026.0134.85
SFT (Router Only)Backbone=Qwen32026.0133.15
SFTBackbone=GPT-OSS2026.0124.83
kNN-MoEBackbone=GPT-OSS2026.0124.35
Zero-shotBackbone=GPT-OSS2026.0123.52
5-shotBackbone=GPT-OSS2026.0119.95
SFT (Router Only)Backbone=GPT-OSS2026.0118.83
SFTBackbone=OLMoE2026.0113.67
kNN-MoEBackbone=OLMoE2026.0113.27
Zero-shotBackbone=OLMoE2026.0113.02
SFT (Router Only)Backbone=OLMoE2026.0111.66
5-shotBackbone=OLMoE2026.0111.09