Mathematical Reasoning on Minerva Math (Score)
42.6ScoreQwen3-4B-GRLO
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-4B-GRLOBackbone=Qwen3-4B, Training=GRLO2026.05 | 42.6 | |
| Qwen3-4B-GRLO+RLVRBackbone=Qwen3-4B, Training=GRLO+RLVR2026.05 | 41.5 | |
| General-Reasoner-4BBackbone=General-Reasoner-4B2026.05 | 25.7 | |
| Qwen3-4B-BaseBackbone=Qwen3-4B2026.05 | 20.6 |