Malicious Package Detection on PyPI Extensive Imports (test) benchmark leaderboard · SOTA2 Research

Malicious Package Detection on PyPI Extensive Imports (test)

95AccuracyH2GLM

Evaluation Results

Method	Links
H2GLM2026.07		95	95.51	94.44	94.97
HERCULE2026.07		82.22	75.89	94.44	84.16
Warehouse2026.07		67.22	91.89	37.78	53.54
Qwen3-14b2026.07		49.45	49.18	32.97	39.46
Clustering2026.07		46.81	44.44	25.53	32.43
MPHunter2026.07		40.56	41.24	44.44	42.78
Llama3-8b2026.07		34.07	34.41	35.16	34.78
DeepSeek-R1-70b2026.07		31.87	18.87	10.99	13.89
Bandit4Mal2026.07		30.56	29.41	27.78	28.57
Naive Bayes2026.07		29.79	22.86	17.02	19.51
Qwen2.5-7b2026.07		29.67	19.67	13.19	15.79
OSSGadget2026.07		29.44	17.78	8.89	11.85