Mathematical Problem Solving on MATH 500 (Pass@k)
79.3Pass@kGCPO
Evaluation Results
| Method | Links | |
|---|---|---|
| GCPOk=64, Base Model=Qwen3-1.7B2026.05 | 79.3 | |
| GCPOk=32, Base Model=Qwen3-1.7B2026.05 | 78.4 | |
| DQOk=64, Base Model=Qwen3-1.7B2026.05 | 78.2 | |
| DQOk=32, Base Model=Qwen3-1.7B2026.05 | 78 | |
| GRPOk=64, Base Model=Qwen3-1.7B2026.05 | 77.6 | |
| GRPOk=32, Base Model=Qwen3-1.7B2026.05 | 77.3 | |
| GCPOk=16, Base Model=Qwen3-1.7B2026.05 | 77 | |
| Basek=64, Base Model=Qwen3-1.7B2026.05 | 76.9 | |
| DQOk=16, Base Model=Qwen3-1.7B2026.05 | 76.3 | |
| Basek=32, Base Model=Qwen3-1.7B2026.05 | 75.8 | |
| GCPOk=8, Base Model=Qwen3-1.7B2026.05 | 75.6 | |
| GRPOk=16, Base Model=Qwen3-1.7B2026.05 | 75.4 | |
| DQOk=8, Base Model=Qwen3-1.7B2026.05 | 74.8 | |
| GRPOk=8, Base Model=Qwen3-1.7B2026.05 | 74 | |
| Basek=16, Base Model=Qwen3-1.7B2026.05 | 73.9 | |
| GCPOk=4, Base Model=Qwen3-1.7B2026.05 | 73.5 | |
| DQOk=4, Base Model=Qwen3-1.7B2026.05 | 72.7 | |
| GRPOk=4, Base Model=Qwen3-1.7B2026.05 | 71.8 | |
| Basek=8, Base Model=Qwen3-1.7B2026.05 | 71.1 | |
| GCPOk=2, Base Model=Qwen3-1.7B2026.05 | 70.3 | |
| DQOk=2, Base Model=Qwen3-1.7B2026.05 | 69.4 | |
| GRPOk=2, Base Model=Qwen3-1.7B2026.05 | 68.2 | |
| Basek=4, Base Model=Qwen3-1.7B2026.05 | 67.6 | |
| Basek=2, Base Model=Qwen3-1.7B2026.05 | 62.5 |