Mathematical Reasoning on HMMT25 (Acc., Rounds, #Samples)
33.6AccuracySC@32
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SC@32Model=Qwen3-4B-Instruct-Thinking2026.06 | 33.6 | 1 | 32 | |
| ASCModel=Qwen3-4B-Instruct-Thinking2026.06 | 33.6 | 17 | 17 | |
| ESCModel=Qwen3-4B-Instruct-Thinking2026.06 | 33.6 | 4.4 | 22 | |
| RL-guidedModel=Qwen3-4B-Instruct-Thinking2026.06 | 33.6 | 3 | 11.8 | |
| RL-guidedModel=Qwen3-1.7B-Thinking2026.06 | 26.7 | 3.5 | 14 | |
| ESCModel=Qwen3-1.7B-Thinking2026.06 | 26.4 | 4.9 | 24.6 | |
| ASCModel=Qwen3-1.7B-Thinking2026.06 | 26 | 20.2 | 20.2 | |
| SC@32Model=Qwen3-1.7B-Thinking2026.06 | 25.9 | 1 | 32 | |
| SC@32Model=Qwen3-0.6B-Thinking2026.06 | 16.4 | 1 | 32 | |
| ASCModel=Qwen3-0.6B-Thinking2026.06 | 16.4 | 23.2 | 23.2 | |
| ESCModel=Qwen3-0.6B-Thinking2026.06 | 16.4 | 5.9 | 29.7 | |
| RL-guidedModel=Qwen3-0.6B-Thinking2026.06 | 16.4 | 4.5 | 18 | |
| SC@32Model=GPT-4.1-nano2026.06 | 12.7 | 1 | 32 | |
| ASCModel=GPT-4.1-nano2026.06 | 12.7 | 23.7 | 23.7 | |
| RL-guidedModel=GPT-4.1-nano2026.06 | 12.6 | 7.3 | 18.6 | |
| ESCModel=GPT-4.1-nano2026.06 | 12.3 | 5.6 | 28.2 |