List your company

Mathematics on AIME25 (Avg@16)

69.2Avg@16TRD
26.996837.953448.9159.8666May 8, 2026May 13, 2026May 18, 2026May 23, 2026May 28, 2026Jun 2, 2026Jun 7, 2026

Evaluation results

MethodLinks
TRDBackbone=Qwen3-8B (w/ thinking), Traj.=yr2026.0669.2
+Reverse KL w/ Top-KBackbone=Qwen3-8B (w/ thinking), Traj.=yo2026.0668.5
+Forward KL w/ ClipBackbone=Qwen3-8B (w/ thinking), Traj.=yo2026.0668.3
+Reverse KLBackbone=Qwen3-8B (w/ thinking), Traj.=yo2026.0668.2
BaseBackbone=Qwen3-8B (w/ thinking), Traj.=–2026.0666.7
+Forward KLBackbone=Qwen3-8B (w/ thinking), Traj.=yo2026.0665.4
TRDBackbone=Qwen3-4B-Instruct-2507, Traj.=yr2026.0649.4
+Forward KL w/ ClipBackbone=Qwen3-4B-Instruct-2507, Traj.=yo2026.0649.2
+Reverse KL w/ Top-KBackbone=Qwen3-4B-Instruct-2507, Traj.=yo2026.0649
+Forward KLBackbone=Qwen3-4B-Instruct-2507, Traj.=yo2026.0648.8
+Reverse KLBackbone=Qwen3-4B-Instruct-2507, Traj.=yo2026.0648.8
BaseBackbone=Qwen3-4B-Instruct-2507, Traj.=–2026.0648.2
AIPOModel=Qwen2.5-32B-Instruct2026.0539.69
GRPOModel=Qwen2.5-32B-Instruct2026.0534.48
SFTModel=Qwen2.5-32B-Instruct2026.0531.26
OriginalModel=Qwen2.5-32B-Instruct2026.0528.62