Mathematical Problem Solving on Olympiad Bench
49.1Pass@8 AccuracyVI-CuRL
Evaluation Results
| Method | Links | |
|---|---|---|
| VI-CuRLModel=Qwen2.5-Math-7B, Verifier Setting=Majority Vote (w/o. Verifier)2026.02 | 49.1 | |
| VI-CuRLModel=Qwen2.5-Math-7B, Verifier Setting=Entropy (w/o. Verifier)2026.02 | 43.4 | |
| No CurriculumModel=Qwen2.5-Math-7B, Verifier Setting=Majority Vote (w/o. Verifier)2026.02 | 41.9 | |
| No CurriculumModel=Qwen2.5-Math-7B, Verifier Setting=Entropy (w/o. Verifier)2026.02 | 35.2 | |
| Base (No RL)Model=Qwen2.5-Math-7B, Verifier Setting=None2026.02 | 33.9 | |
| No CurriculumModel=Llama3.2-3B-Instruct, Verifier Setting=Majority Vote (w/o. Verifier)2026.02 | 24 | |
| VI-CuRLModel=Llama3.2-3B-Instruct, Verifier Setting=Majority Vote (w/o. Verifier)2026.02 | 23.3 | |
| VI-CuRLModel=Llama3.2-3B-Instruct, Verifier Setting=Entropy (w/o. Verifier)2026.02 | 21.8 | |
| Base (No RL)Model=Llama3.2-3B-Instruct, Verifier Setting=None2026.02 | 16.4 | |
| No CurriculumModel=Llama3.2-3B-Instruct, Verifier Setting=Entropy (w/o. Verifier)2026.02 | 1.5 |