List your company

Mathematical Reasoning on AIME (pass@1/mean@32)

65.6Pass@1/Mean@32p*
57.2859.4461.663.76Apr 30, 2026

Evaluation results

MethodLinks
p*Setting=GRPO2026.0465.6
p* smooth(alpha = 0.01)Setting=GRPO2026.0465.3
p* smooth(alpha = 0.01)Setting=ZVF2026.0462.9
No samplingSetting=GRPO2026.0461.3
No samplingSetting=ZVF2026.0458.1
p*Setting=ZVF2026.0457.6