List your company

Mathematics on MATH 500 (avg@32 accuracy)

83.3Accuracy (avg@32)IBTPO
49.91658.58367.2575.917May 27, 2026

Evaluation results

MethodLinks
IBTPOBase Model=Qwen3-8B-Base2026.0583.3
TreeRLBase Model=Qwen3-8B-Base2026.0582.5
TreePOBase Model=Qwen3-8B-Base2026.0582.2
IBROBase Model=Qwen3-8B-Base2026.0582
GRPO w/ Entropy Reg†Base Model=Qwen3-8B-Base2026.0581.8
Vanilla GRPOBase Model=Qwen3-8B-Base2026.0581.5
GRPO w/ Clip-higherBase Model=Qwen3-8B-Base2026.0581.2
IBTPOBase Model=Qwen3-1.7B-Base2026.0570.1
TreeRLBase Model=Qwen3-1.7B-Base2026.0567.2
Vanilla GRPOBase Model=Qwen3-1.7B-Base2026.0566.8
TreePOBase Model=Qwen3-1.7B-Base2026.0566.6
IBROBase Model=Qwen3-1.7B-Base2026.0566.4
GRPO w/ Clip-higherBase Model=Qwen3-1.7B-Base2026.0565.8
Initial ModelBase Model=Qwen3-8B-Base2026.0565.7
GRPO w/ Entropy Reg†Base Model=Qwen3-1.7B-Base2026.0564.8
Initial ModelBase Model=Qwen3-1.7B-Base2026.0551.2