Mathematical Reasoning on Mathematical Reasonings Average
60.9AccuracyQwen3-14B + GRPO
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen3-14B + GRPOBackbone=Qwen3-14B, Training=GRPO2026.05 | 60.9 | 5,024 | |
| Qwen3-14B + SLATBackbone=Qwen3-14B, Training=SLAT2026.05 | 60.6 | 2,677 | |
| Qwen3-14BBackbone=Qwen3-14B2026.05 | 38.8 | 1,216 |