List your company

Math on BeyondAIME (Avg@16)

42.8Avg@16TRD
31.15234.17637.240.224Jun 7, 2026

Evaluation results

MethodLinks
TRDBackbone=Qwen3-8B (w/ thinking), Traj.=yr2026.0642.8
+Reverse KL w/ Top-KBackbone=Qwen3-8B (w/ thinking), Traj.=yo2026.0641.9
BaseBackbone=Qwen3-8B (w/ thinking), Traj.=–2026.0641.6
+Reverse KLBackbone=Qwen3-8B (w/ thinking), Traj.=yo2026.0640.8
+Forward KL w/ ClipBackbone=Qwen3-8B (w/ thinking), Traj.=yo2026.0640.2
+Forward KLBackbone=Qwen3-8B (w/ thinking), Traj.=yo2026.0639.6
TRDBackbone=Qwen3-4B-Instruct-2507, Traj.=yr2026.0632.7
+Forward KL w/ ClipBackbone=Qwen3-4B-Instruct-2507, Traj.=yo2026.0632.4
+Forward KLBackbone=Qwen3-4B-Instruct-2507, Traj.=yo2026.0632.3
+Reverse KL w/ Top-KBackbone=Qwen3-4B-Instruct-2507, Traj.=yo2026.0632.3
BaseBackbone=Qwen3-4B-Instruct-2507, Traj.=–2026.0632.2
+Reverse KLBackbone=Qwen3-4B-Instruct-2507, Traj.=yo2026.0631.6