Mathematical Reasoning on AMC 23 (Pass@1, Pass@16)
65Pass@1APMPO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| APMPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.04 | 65 | 92.5 | |
| APMPOBackbone=Qwen2.5-Math-1.5B-Instruct2026.04 | 62.5 | 85 | |
| GMPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.04 | 62.5 | 87.5 | |
| DAPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.04 | 60 | 90 | |
| DAPOBackbone=Qwen2.5-Math-1.5B-Instruct2026.04 | 57.5 | 80 | |
| GRPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.04 | 57.5 | 82.5 | |
| GMPOBackbone=Qwen2.5-Math-1.5B-Instruct2026.04 | 55 | 82.5 | |
| GRPOBackbone=Qwen2.5-Math-1.5B-Instruct2026.04 | 52.5 | 75 | |
| BaseBackbone=Qwen2.5-Math-1.5B-Instruct2026.04 | 47.5 | 70 | |
| BaseBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.04 | 47.5 | 70 | |
| DAPOBackbone=Qwen2.5-3B-Instruct2026.04 | 45 | 65 | |
| APMPOBackbone=Qwen2.5-3B-Instruct2026.04 | 45 | 70 | |
| GMPOBackbone=Qwen2.5-3B-Instruct2026.04 | 42.5 | 60 | |
| GRPOBackbone=Qwen2.5-3B-Instruct2026.04 | 40 | 60 | |
| BaseBackbone=Qwen2.5-3B-Instruct2026.04 | 35 | 55 |