List your company

Mathematics on AMC 23 (avg@32 accuracy)

57.9Avg@32 AccuracyIBTPO
23.16432.18241.250.218May 27, 2026

Evaluation results

MethodLinks
IBTPOBase Model=Qwen3-8B-Base2026.0557.9
TreeRLBase Model=Qwen3-8B-Base2026.0556.1
TreePOBase Model=Qwen3-8B-Base2026.0555.8
IBROBase Model=Qwen3-8B-Base2026.0555.3
GRPO w/ Entropy Reg†Base Model=Qwen3-8B-Base2026.0554.3
Vanilla GRPOBase Model=Qwen3-8B-Base2026.0553.5
GRPO w/ Clip-higherBase Model=Qwen3-8B-Base2026.0553.3
Initial ModelBase Model=Qwen3-8B-Base2026.0541.7
IBTPOBase Model=Qwen3-1.7B-Base2026.0538.7
TreeRLBase Model=Qwen3-1.7B-Base2026.0537
TreePOBase Model=Qwen3-1.7B-Base2026.0536.5
Vanilla GRPOBase Model=Qwen3-1.7B-Base2026.0535.6
GRPO w/ Clip-higherBase Model=Qwen3-1.7B-Base2026.0535
IBROBase Model=Qwen3-1.7B-Base2026.0534.6
GRPO w/ Entropy Reg†Base Model=Qwen3-1.7B-Base2026.0533.6
Initial ModelBase Model=Qwen3-1.7B-Base2026.0524.5