Mathematical Reasoning on AIME 2024 (Accuracy and Token Costs)
40Sample CountSC
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SCBackbone=Qwen2.5-3B2026.02 | 40 | — | 42.3 | 0.112 | |
| SCBackbone=Qwen2.5-7B2026.02 | 40 | — | 42.2 | 0.1667 | |
| SCBackbone=Gemma3-4B2026.02 | 40 | — | 163.8 | 0.1333 | |
| ESCBackbone=Qwen2.5-3B2026.02 | 38.81 | — | 43 | 0.112 | |
| DSCBackbone=Qwen2.5-3B2026.02 | 37.85 | 4.9 | 41.3 | 0.112 | |
| ESCBackbone=Qwen2.5-7B2026.02 | 37.11 | — | 41.5 | 0.1667 | |
| ACBackbone=Qwen2.5-3B2026.02 | 34.68 | — | 44.1 | 0.112 | |
| DSCBackbone=Qwen2.5-7B2026.02 | 34.11 | 5 | 37.8 | 0.1667 | |
| ESCBackbone=Gemma3-4B2026.02 | 33.7 | — | 139.3 | 0.1333 | |
| ACBackbone=Qwen2.5-7B2026.02 | 32.3 | — | 41.2 | 0.1667 | |
| DSCBackbone=Gemma3-4B2026.02 | 29.78 | 17 | 122.5 | 0.1333 | |
| ACTSCBackbone=Qwen2.5-3B2026.02 | 28.17 | — | 33.7 | 0.112 | |
| ACBackbone=Gemma3-4B2026.02 | 27.3 | — | 116.7 | 0.1333 | |
| ACTSCBackbone=Qwen2.5-7B2026.02 | 24.7 | — | 29 | 0.1667 | |
| ACTSCBackbone=Gemma3-4B2026.02 | 18.8 | — | 78.1 | 0.1 |