STRING on AB-DB benchmark leaderboard · SOTA2 Research

STRING on AB-DB

28.7LLM Judge AccuracyA3

Evaluation Results

Method	Links
A32026.05		28.7
A32026.05		27.8
RetroAgent2026.05		16.3
ReACT2026.05		16
GSPO2026.05		14.7
HGPO2026.05		14.7
GiGPO2026.05		10.7
LATS2026.05		9.9
rStar2026.05		5.6