Math Reasoning on AIME 2024 (Final Accuracy, Avg Accuracy, Anytime Index)
100Final AccuracyPDP
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| PDPModel=Grok-3-mini2026.01 | 100 | 86 | 87.1 | |
| LEAPModel=Grok-3-mini2026.01 | 86.7 | 77.9 | 81.7 | |
| PDP(+)Model=Grok-3-mini2026.01 | 83.3 | 79 | 82 | |
| BaseModel=Grok-3-mini2026.01 | 80.6 | 75.2 | 80.9 | |
| PDPModel=GPT-oss-120B2026.01 | 38.9 | 41.7 | 52.9 | |
| LEAPModel=Llama-3.3-70B2026.01 | 33.7 | 21.7 | 26.3 | |
| BaseModel=GPT-oss-120B2026.01 | 32 | 43.9 | 54.4 | |
| LEAPModel=GPT-oss-120B2026.01 | 30.2 | 40.5 | 49.8 | |
| PDP(+)Model=GPT-oss-120B2026.01 | 30.2 | 38.8 | 46.5 | |
| PDP(+)Model=Grok-32026.01 | 25 | 11.8 | 11.5 | |
| PDPModel=Grok-32026.01 | 24.9 | 12.6 | 12.3 | |
| PDP(+)Model=Llama-3.3-70B2026.01 | 24.1 | 24.6 | 28.6 | |
| BaseModel=Grok-32026.01 | 24 | 11.1 | 11 | |
| LEAPModel=Grok-32026.01 | 22.8 | 12 | 11.9 | |
| PDPModel=Llama-3.3-70B2026.01 | 22.3 | 25.3 | 29 | |
| PDPModel=GPT-oss-20B2026.01 | 17.7 | 29.1 | 40.6 | |
| BaseModel=Llama-3.3-70B2026.01 | 17.1 | 23.6 | 28.7 | |
| BaseModel=GPT-oss-20B2026.01 | 16.5 | 22.7 | 40.4 | |
| PDP(+)Model=GPT-oss-20B2026.01 | 13.9 | 28.9 | 38.6 | |
| LEAPModel=GPT-4.12026.01 | 10.2 | 7.5 | 9.5 | |
| LEAPModel=GPT-oss-20B2026.01 | 9.8 | 25.5 | 33.7 | |
| PDPModel=GPT-4.12026.01 | 7.7 | 3.9 | 5.2 | |
| PDPModel=GPT-4o2026.01 | 5.1 | 3.1 | 4.3 | |
| BaseModel=GPT-4o2026.01 | 3.8 | 2.9 | 4.3 | |
| LEAPModel=GPT-4o2026.01 | 3.7 | 3 | 5 | |
| BaseModel=GPT-4.12026.01 | 2.8 | 4.3 | 7.2 | |
| PDP(+)Model=GPT-4o2026.01 | 2.4 | 2.8 | 4.4 | |
| PDP(+)Model=GPT-4.12026.01 | 1.2 | 3.4 | 6.6 |