List your company

Mathematical Problem Solving on AIME, AMC, MATH, GSM8K (test)

58.5AIME Accuracyp*
58.29258.34658.458.454Apr 30, 2026

Evaluation results

MethodLinks
p*Setting=GRPO, Backbone=8B-Base, Evaluation Protocol=pass@1/mean@322026.0458.574.786.795.278.8
p* smooth(alpha = 0.01)Setting=GRPO, Backbone=8B-Base, Evaluation Protocol=pass@1/mean@322026.0458.57486.794.978.5
No samplingSetting=GRPO, Backbone=8B-Base, Evaluation Protocol=pass@1/mean@322026.0458.372.8879678.5