List your company

Eventual-success tracking on MATH-500 Level 5

26.1AUROCSBBT
15.59618.32321.0523.777May 26, 2026

Evaluation results

MethodLinks
SBBTModel=7B, Signal=hybrid, Reference=temporal baseline2026.0526.1—
SBBTModel=Qwen2.5, Signal=hybrid, Reference=temporal baseline2026.0519—
SBBTModel=7B, Signal=hybrid, Reference=EMA2026.0516—
SBBTModel=Qwen3-14B, Signal=score, Reference=calibrated last-step2026.05—0.061