Share your thoughts, 1 month free Claude Pro on usSee more

Mathematical Reasoning on AMC23 (Pass@k)

98.6Pass@k

Qwen2.5-7B + PPO

Updated 26d ago

Evaluation Results

Method	Links
Qwen2.5-7B + PPO 2025.09		98.6
Qwen2.5-7B 2025.09		98.4
Qwen2.5-7B + GRPO w/ VERL. 2025.09		98.3
Qwen2.5-7B + PPO w/ VERL. 2025.09		98
Qwen2.5-7B + GRPO 2025.09		97.8
Mathstral-7B-v0.1 + GRPO w/ VERL. 2025.09		97
Qwen2.5-3B + PPO w/ VERL. 2025.09		96.8
Qwen2.5-3B 2025.09		96.7
Qwen2.5-3B + PPO 2025.09		96.5
Qwen2.5-3B + GRPO w/ VERL. 2025.09		95.9
Llama-3.2-3B-Instruct + GRPO w/ VERL. 2025.09		95.7
Llama-3.2-3B-Instruct + GRPO 2025.09		95.4
Llama-3.1-8B-Instruct + GRPO w/ VERL. 2025.09		95.1
Llama-3.1-8B-Instruct + GRPO 2025.09		94.9
Llama-3.2-3B-Instruct + PPO w/ VERL. 2025.09		94.7
Llama-3.1-8B-Instruct 2025.09		94.6
Llama-3.2-3B-Instruct + PPO 2025.09		94.5
Mathstral-7B-v0.1 + PPO w/ VERL. 2025.09		93.8
Llama-3.2-3B-Instruct 2025.09		93.5
Llama-3.1-8B-Instruct + PPO 2025.09		92.4
Qwen2.5-3B + GRPO 2025.09		92.2
Llama-3.1-8B-Instruct + PPO w/ VERL. 2025.09		91.9
Mathstral-7B-v0.1 + PPO 2025.09		91.7
Mathstral-7B-v0.1 2025.09		88.5
Mathstral-7B-v0.1 + GRPO 2025.09		87.3
Mistral-7B-v0.3 2025.09		73.5
Mistral-7B-v0.3 + GRPO w/ VERL. 2025.09		64.5
Mistral-7B-v0.3 + GRPO 2025.09		63.2
GRPO(λ)-0.99 2026.05		60
GRPO 2026.05		57.5
GRPO(λ)-0.9 2026.05		57.5
S-trace-0.9 2026.05		57.5
OPO 2026.05		55
Mistral-7B-v0.3 + PPO w/ VERL. 2025.09		46.5
Mistral-7B-v0.3 + PPO 2025.09		46.4