CVE-2026-41523

EUVD-2026-38406
vLLM is an inference and serving engine for large language models (LLMs). Prior to 0.22.0, an assert-based security check in vLLM's activation function loading allows any unauthenticated attacker to achieve arbitrary code execution on the server by publishing a malicious HuggingFace model, when vLLM runs in Python optimized mode (python -O or PYTHONOPTIMIZE=1). This vulnerability is fixed in 0.22.0.
Code Injection
ProviderTypeBase ScoreAtk. VectorAtk. ComplexityPriv. RequiredVector
NISTPrimary
7.5 HIGH
NETWORK
HIGH
NONE
CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H
Base Score
CVSS 3.x
EPSS Score
Percentile: 51.34%
Affected Products (NVD)
VendorProductVersion
vllmvllm
𝑥
< 0.22.0
𝑥
= Vulnerable software versions
Early Detection
Affected products identified ahead of NVD analysis through intelligence sources.
VendorProductVersionSource
Red HatRed Hat AI Inference Server 3.2
1782951012 ≤
𝑥
< *
ADP
Red HatRed Hat AI Inference Server 3.2
1782951244 ≤
𝑥
< *
ADP