STRING on AB-OS benchmark leaderboard · SOTA2 Research

STRING on AB-OS

61.4LLM Judge AccuracyGSPO

Evaluation Results

Method	Links
GSPO2026.05		61.4
A32026.05		60.7
A32026.05		58.6
ReACT2026.05		57.9
GiGPO2026.05		54.5
RetroAgent2026.05		49
HGPO2026.05		46.9
LATS2026.05		40
rStar2026.05		35.9