Mathematical Reasoning on Average (MATH500, AMC, AIME24, GPQA)
43.8Acc @ FirstCoVerRL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CoVerRLModel=Qwen2.5-7B2026.03 | 43.8 | 45.8 | |
| TTRLModel=Qwen2.5-7B2026.03 | 41.1 | 41.1 | |
| CoVerRLModel=Qwen3-1.7B-Base2026.03 | 36.9 | 38.7 | |
| CoVerRLModel=Llama-3.2-3B-Instruct2026.03 | 33.2 | 35.3 | |
| TTRLModel=Qwen3-1.7B-Base2026.03 | 33.1 | 33 | |
| TTRLModel=Llama-3.2-3B-Instruct2026.03 | 30 | 29.4 | |
| Base ModelModel=Qwen2.5-7B2026.03 | 27.7 | 28.5 | |
| Base ModelModel=Qwen3-1.7B-Base2026.03 | 27.4 | 27.1 | |
| Base ModelModel=Llama-3.2-3B-Instruct2026.03 | 22.8 | 22 |