List your company

Math Reasoning on MATH 500 (Score %)

90.2MATH 500 ScoreSoft-RLVR
87.49688.19888.989.602May 27, 2026

Evaluation results

MethodLinks
Soft-RLVRVerifier=GPT-OSS-120B, Checklist source=DeepSeek R1, Training step=5002026.0590.2
Soft-RLVRVerifier=GPT-OSS-20B, Checklist source=DeepSeek R1, Training step=5002026.0589.6
Oracle rule-basedReward=Oracle rule-based, Training step=5002026.0587.8
Command R7B (No RL)Reward=None (Initial Policy), Training step=5002026.0587.6