List your company
ResearchFormerly WizWand

Mathematical Reasoning on MathVision (out-of-distribution)

27.93AccuracyCalibRL
23.177224.411125.64526.8789Feb 22, 2026
Updated 8mo ago

Evaluation results

MethodLinks
CalibRL2026.0227.93
GRPO2026.0227.89
RL-PLUS2026.0227.14
DAPO2026.0227.07
LUFFY2026.0226.64
SFT+GRPO2026.0223.36