Mathematical Problem Solving on AIME 2024 (Accuracy (avg@k))
80.21Accuracy (avg@k)TRACE (FKL on Ky)
Evaluation Results
| Method | Links | |
|---|---|---|
| TRACE (FKL on Ky)Backbone=Qwen3-8B, Symmetric Think evaluation=true2026.05 | 80.21 | |
| TRACE (RKL on Ey)Backbone=Qwen3-8B, Symmetric Think evaluation=true2026.05 | 78.54 | |
| GRPOBackbone=Qwen3-8B, Symmetric Think evaluation=true2026.05 | 77.08 | |
| BaseBackbone=Qwen3-8B, Symmetric Think evaluation=true2026.05 | 76.25 | |
| RLSDBackbone=Qwen3-8B, Symmetric Think evaluation=true2026.05 | 75.42 | |
| SRPOBackbone=Qwen3-8B, Symmetric Think evaluation=true2026.05 | 69.17 | |
| SDPOBackbone=Qwen3-8B, Symmetric Think evaluation=true2026.05 | 52.5 | |
| RLSDBackbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.05 | 45.42 | |
| TRACE (RKL on Ey)Backbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.05 | 44.92 | |
| TRACE (FKL on Ky)Backbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.05 | 44.83 | |
| GRPOBackbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.05 | 44.58 | |
| BaseBackbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.05 | 43.75 | |
| SRPOBackbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.05 | 24.17 | |
| SDPOBackbone=Qwen3-1.7B, Symmetric Think evaluation=true2026.05 | 16.25 |