Multiple-choice Question Answering on PIQA (test) benchmark leaderboard · SOTA2 Research

Multiple-choice Question Answering on PIQA (test)

77.8Accuracy (PIQA)Oracle

Evaluation Results

Method	Links
Oracle2026.04		77.8
REALM2026.04		77.67
REALM2026.04		77.06
REALM2026.04		76.45
NOISY2026.04		75.6
Oracle2026.04		67.68
REALM2026.04		66.92
REALM2026.04		66.22
REALM2026.04		65.7
NOISY2026.04		64.87
AR Transformer2026.02		63.71
MDLM2026.02		57.89
REALM2026.04		54.41
Oracle2026.04		52.88
REALM2026.04		52.14
Duo++2026.02		52.12
Duo++2026.02		51.25
Duo2026.02		51.14
NOISY2026.04		51.08
REALM2026.04		51.06
NOISY2026.04		50.84
Duo++2026.02		50.65
NOISY2026.04		50.16
NOISY2026.04		46.31
NOISY2026.04		46.31
NOISY2026.04		27.62
NOISY2026.04		26.27