Mathematical Reasoning on OlympiadBench (Score)
45.2ScoreGeneral-Reasoner-4B
Evaluation Results
| Method | Links | |
|---|---|---|
| General-Reasoner-4BBackbone=General-Reasoner-4B2026.05 | 45.2 | |
| Qwen3-4B-GRLOBackbone=Qwen3-4B, Training=GRLO2026.05 | 43.4 | |
| Qwen3-4B-GRLO+RLVRBackbone=Qwen3-4B, Training=GRLO+RLVR2026.05 | 42.7 | |
| Qwen3-4B-BaseBackbone=Qwen3-4B2026.05 | 34.2 |