List your company

Physical Reasoning on PIQA

79.71PIQA Normalized PerformanceFP16
13.243630.499347.75565.0107Dec 16, 2025Jan 9, 2026Feb 2, 2026Feb 27, 2026Mar 23, 2026Apr 16, 2026May 11, 2026

Evaluation results

MethodLinks
FP16Bit-width=FP16, Evaluation Protocol=zero-shot2026.0579.71
QTIP 2-bit PTQBit-width=2-bit, Evaluation Protocol=zero-shot2026.0577.97
BCJR-QAT-N4Bit-width=2 bpw, Evaluation Protocol=zero-shot, Configuration=N42026.0576.93
Dense 1.3BArchitecture=Dense, Total Parameters=1.3B, Precision=bf16, Shot count=0-shot, Evaluation Framework=lm-eval-harness 0.4.72026.0565.56
MoL 0.61B/2.08BArchitecture=MoL, Total Parameters=2.08B, Active Parameters=0.61B, Precision=bf16, Shot count=0-shot, Evaluation Framework=lm-eval-harness 0.4.72026.0564.09
Dense 0.7BArchitecture=Dense, Total Parameters=0.7B, Precision=bf16, Shot count=0-shot, Evaluation Framework=lm-eval-harness 0.4.72026.0563.98
DualAlpha (α)=63/64, Data repetitions=12025.1240.9
AutoregressiveAlpha (α)=1, Data repetitions=12025.1239.4
DualAlpha (α)=3/4, Data repetitions=322025.1236.1
AutoregressiveAlpha (α)=1, Data repetitions=322025.1233.3
DualAlpha (α)=1/8, Data repetitions=1282025.1228.1
AutoregressiveAlpha (α)=1, Data repetitions=1282025.1215.8