Mathematical Reasoning on OlympiadBench (val)
54.16AccuracyGenAC
Evaluation results
| Method | Links | |
|---|---|---|
| GenACAlgorithm=GenAC, Base Model=Qwen3-8B-Base, Sampled generations=16, RL Setup=Zero RL2026.04 | 54.16 | |
| RLOOAlgorithm=RLOO, Base Model=Qwen3-8B-Base, Sampled generations=16, RL Setup=Zero RL2026.04 | 53.94 | |
| VC-PPOAlgorithm=VC-PPO, Base Model=Qwen3-8B-Base, Sampled generations=16, RL Setup=Zero RL2026.04 | 53.5 | |
| GRPOAlgorithm=GRPO, Base Model=Qwen3-8B-Base, Sampled generations=16, RL Setup=Zero RL2026.04 | 52.92 |