Mathematical Reasoning on AIME 2023-2025 (pass@4)
11.11pass@4 AccuracySAGC
Evaluation Results
| Method | Links | |
|---|---|---|
| SAGCBackbone=Llama3.2-3b, Algorithm=DAPO, Group Size=Dynamic2026.06 | 11.11 | |
| SAGCBackbone=Qwen2.5-3b, Algorithm=DAPO, Group Size=Dynamic2026.06 | 10 | |
| OpenRLHFBackbone=Llama3.2-3b, Algorithm=DAPO, Group Size=162026.06 | 10 | |
| OpenRLHFBackbone=Qwen2.5-3b, Algorithm=DAPO, Group Size=162026.06 | 8.89 | |
| OpenRLHFBackbone=Llama3.2-3b, Algorithm=GRPO, Group Size=162026.06 | 8.89 | |
| OpenRLHFBackbone=Qwen2.5-3b, Algorithm=GRPO, Group Size=162026.06 | 7.78 | |
| SAGCBackbone=Qwen2.5-3b, Algorithm=GRPO, Group Size=Dynamic2026.06 | 7.78 | |
| OpenRLHFBackbone=Llama3.2-3b, Algorithm=GRPO, Group Size=82026.06 | 7.78 | |
| SAGCBackbone=Llama3.2-3b, Algorithm=GRPO, Group Size=Dynamic2026.06 | 7.78 | |
| OpenRLHFBackbone=Qwen2.5-3b, Algorithm=GRPO, Group Size=42026.06 | 6.67 | |
| OpenRLHFBackbone=Qwen2.5-3b, Algorithm=GRPO, Group Size=82026.06 | 6.67 | |
| OpenRLHFBackbone=Llama3.2-3b, Algorithm=GRPO, Group Size=42026.06 | 6.67 | |
| OpenRLHFBackbone=Llama3.2-3b, Algorithm=DAPO, Group Size=42026.06 | 6.67 | |
| OpenRLHFBackbone=Llama3.2-3b, Algorithm=DAPO, Group Size=82026.06 | 6.67 | |
| OpenRLHFBackbone=Qwen2.5-3b, Algorithm=DAPO, Group Size=42026.06 | 5.56 | |
| OpenRLHFBackbone=Qwen2.5-3b, Algorithm=DAPO, Group Size=82026.06 | 5.56 |