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

Code Reasoning on CRUXEval

76.87Accuracy

PPoT + Qwen2.5 Coder

Updated 7d ago

Evaluation Results

Method	Links
PPoT + Qwen2.5 Coder 2026.04		76.87
Qwen2.5 Coder 2026.04		73.37
Qwen2.5-Math-72B-Instruct 2025.02		68.6
QwQ-32B-Preview* 2025.02		65.2
PPoT + Qwen2.5 Coder 2026.04		63.87
PPoT + Qwen2.5 Coder 2026.04		60.62
Llama-3.1-70B-Instruct* 2025.02		59.6
Qwen2.5 Coder 2026.04		59.13
Qwen2.5 Coder 2026.04		57.12
PPoT + Qwen2.5 Coder 2026.04		53.13
DataFlow-Code-10K 2025.12		52.9
Self-OSS 2025.12		52.6
DataFlow-Code-5K 2025.12		52.5
DataFlow-Code-1K 2025.12		51
Qwen2.5-Math-7B-S2R-ORL 2025.02		50.9
Code Alpaca-1K 2025.12		50.2
Eurus-2-7B-PRIME 2025.02		50
DataFlow-Code-10K 2025.12		48.8
DataFlow-Code-5K 2025.12		48.6
Qwen2.5-Math-7B-S2R-BI 2025.02		48
DataFlow-Code-1K 2025.12		48
Qwen2.5-14B-Instruct 2025.12		48
Self-OSS 2025.12		46.9
Qwen2.5 Coder 2026.04		46.13
PPoT + Qwen2.5 Coder 2026.04		45.87
Code Alpaca-1K 2025.12		45.6
Qwen2.5-7B-Instruct 2025.12		44.8
Qwen2.5-Math-7B 2025.02		40.8
Qwen2.5 Coder 2026.04		40.5
PPoT + Qwen2.5 Coder 2026.04		37.37
AdamW 2026.05		35.63
OpenMath2-Llama3.1-70B* 2025.02		35.1
MONA 2026.05		35
Muon 2026.05		33.62
Qwen2.5 Coder 2026.04		30.87
Qwen2.5-Math-7B-Instruct 2025.02		28