Mathematics on AIME 24
83.7Avg@32GLM 4.6
Evaluation Results
| Method | Links | |
|---|---|---|
| GLM 4.6Evaluation Mode=Chat2025.12 | 83.7 | |
| LongCat-Flash Exp-ChatEvaluation Mode=Chat2025.12 | 83.1 | |
| DeepSeek V3.2Evaluation Mode=Chat2025.12 | 73.9 | |
| LongCat-Flash ChatEvaluation Mode=Chat2025.12 | 70.4 | |
| IBTPOBase Model=Qwen3-8B-Base2026.05 | 0.197 | |
| TreeRLBase Model=Qwen3-8B-Base2026.05 | 0.178 | |
| TreePOBase Model=Qwen3-8B-Base2026.05 | 0.175 | |
| Vanilla GRPOBase Model=Qwen3-8B-Base2026.05 | 0.171 | |
| GRPO w/ Entropy Reg†Base Model=Qwen3-8B-Base2026.05 | 0.17 | |
| IBROBase Model=Qwen3-8B-Base2026.05 | 0.169 | |
| GRPO w/ Clip-higherBase Model=Qwen3-8B-Base2026.05 | 0.168 | |
| Initial ModelBase Model=Qwen3-8B-Base2026.05 | 0.115 | |
| IBTPOBase Model=Qwen3-1.7B-Base2026.05 | 0.095 | |
| TreeRLBase Model=Qwen3-1.7B-Base2026.05 | 0.079 | |
| GRPO w/ Clip-higherBase Model=Qwen3-1.7B-Base2026.05 | 0.077 | |
| TreePOBase Model=Qwen3-1.7B-Base2026.05 | 0.072 | |
| Vanilla GRPOBase Model=Qwen3-1.7B-Base2026.05 | 0.07 | |
| GRPO w/ Entropy Reg†Base Model=Qwen3-1.7B-Base2026.05 | 0.066 | |
| IBROBase Model=Qwen3-1.7B-Base2026.05 | 0.056 | |
| Initial ModelBase Model=Qwen3-1.7B-Base2026.05 | 0.038 |