Competition Mathematics on AIME 2025 (Acc@t1, Delta Metrics)
80Acc@t1SELF-THOUGHT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SELF-THOUGHTModel=O3-MINI2026.01 | 80 | 6.67 | 0 | |
| REFLEXModel=O3-MINI2026.01 | 76.67 | 3.33 | 0 | |
| INITIAL (t = 0)Model=O3-MINI2026.01 | 73.33 | — | — | |
| SELF-REFINEModel=O3-MINI2026.01 | 73.33 | 6.67 | 6.67 | |
| SELF-THOUGHTModel=DEEPSEEK-R12026.01 | 73.33 | 13.33 | 3.33 | |
| SELF-REFINEModel=DEEPSEEK-R12026.01 | 70 | 20 | 13.33 | |
| SELF-TICKModel=O3-MINI2026.01 | 66.67 | 10 | 16.67 | |
| REFLEXIONModel=O3-MINI2026.01 | 66.67 | 3.33 | 10 | |
| INITIAL (t = 0)Model=DEEPSEEK-R12026.01 | 63.33 | — | — | |
| REFLEXModel=DEEPSEEK-R12026.01 | 63.33 | 6.67 | 6.67 | |
| REFLEXIONModel=DEEPSEEK-R12026.01 | 60 | 6.67 | 10 | |
| SELF-TICKModel=DEEPSEEK-R12026.01 | 53.33 | 6.67 | 16.67 | |
| SELF-THOUGHTModel=GPT-4O-MINI2026.01 | 20 | 13.33 | 0 | |
| SELF-REFINEModel=GPT-4O-MINI2026.01 | 16.67 | 10 | 0 | |
| SELF-THOUGHTModel=GPT-4O2026.01 | 16.67 | 10 | 3.33 | |
| SELF-TICKModel=GPT-4O-MINI2026.01 | 13.33 | 6.67 | 0 | |
| REFLEXModel=GPT-4O-MINI2026.01 | 10 | 3.33 | 0 | |
| INITIAL (t = 0)Model=GPT-4O2026.01 | 10 | — | — | |
| SELF-REFINEModel=GPT-4O2026.01 | 10 | 6.67 | 6.67 | |
| SELF-TICKModel=GPT-4O2026.01 | 10 | 6.67 | 6.67 | |
| REFLEXIONModel=GPT-4O2026.01 | 10 | 6.67 | 6.67 | |
| INITIAL (t = 0)Model=GPT-4O-MINI2026.01 | 6.67 | — | — | |
| REFLEXIONModel=GPT-4O-MINI2026.01 | 6.67 | 3.33 | 3.33 | |
| REFLEXModel=GPT-4O2026.01 | 6.67 | 0 | 3.33 |