List your company

Scientific Reasoning on GPQA Diamond (average@32)

55.19Average@32GRPO
11.56222.888534.21545.5415May 8, 2026

Evaluation results

MethodLinks
GRPOParams=4B2026.0555.19
SODParams=1.7B2026.0538.72
OPSDgtParams=1.7B2026.0535.02
GRPOParams=1.7B2026.0533.55
OPSDhintParams=1.7B2026.0533.46
OPDParams=1.7B2026.0531.73
SFTParams=1.7B2026.0529.85
VanillaParams=1.7B2026.0526.8
SODParams=0.6B2026.0522.19
GRPOParams=0.6B2026.0520.38
OPDParams=0.6B2026.0517.76
OPSDgtParams=0.6B2026.0517.32
OPSDhintParams=0.6B2026.0515.98
SFTParams=0.6B2026.0515.2
VanillaParams=0.6B2026.0513.24