List your company

Agentic Task Solving on AB-OS

79.7Pass@3ReACT
52.03659.21866.473.582May 8, 2026

Evaluation results

MethodLinks
ReACT2026.0579.786.2
GSPO2026.057982.8
A³variant=σ-Reveal2026.0576.279.3
A³variant=Vanilla2026.0572.172.4
GiGPO2026.0567.672.4
HGPO2026.0562.869
RetroAgent2026.0562.869
LATS2026.0559.365.5
rStar2026.0553.158.6