Mathematical Reasoning on Gaokao MathCloze
69.5Pass@1Qwen2.5-7B + GRPO w/ VERL.
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-7B + GRPO w/ VERL.Base Model=Qwen2.5-7B, RL Framework=GRPO, VERL Integration=true2025.09 | 69.5 | |
| Qwen2.5-7B + PPO w/ VERL.Base Model=Qwen2.5-7B, RL Framework=PPO, VERL Integration=true2025.09 | 66.9 | |
| Qwen2.5-7B + GRPOBase Model=Qwen2.5-7B, RL Framework=GRPO, VERL Integration=false2025.09 | 64.4 | |
| Qwen2.5-7B + PPOBase Model=Qwen2.5-7B, RL Framework=PPO, VERL Integration=false2025.09 | 63.6 | |
| Qwen2.5-7BBase Model=Qwen2.5-7B, RL Framework=None, VERL Integration=false2025.09 | 49.2 | |
| Llama-3.2-3B-Instruct + PPO w/ VERL.Base Model=Llama-3.2-3B-Instruct, RL Framework=PPO, VERL Integration=true2025.09 | 27.1 | |
| Llama-3.2-3B-Instruct + GRPOBase Model=Llama-3.2-3B-Instruct, RL Framework=GRPO, VERL Integration=false2025.09 | 23.7 | |
| Llama-3.2-3B-Instruct + GRPO w/ VERL.Base Model=Llama-3.2-3B-Instruct, RL Framework=GRPO, VERL Integration=true2025.09 | 22.9 | |
| Llama-3.2-3B-Instruct + PPOBase Model=Llama-3.2-3B-Instruct, RL Framework=PPO, VERL Integration=false2025.09 | 20.3 | |
| Llama-3.2-3B-InstructBase Model=Llama-3.2-3B-Instruct, RL Framework=None, VERL Integration=false2025.09 | 6.8 |