Back to overview

CVE-2026-41523

HIGH
7.5
CVSS 3.1
Description
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.

Metadata

CVE ID
CVE-2026-41523
State
PUBLISHED
Assigner
GitHub_M
Reserved
2026-04-20 18:18 UTC
Published
2026-06-22 22:18 UTC
Last updated
2026-06-22 22:18 UTC
Primary CWE
CWE-94
CWE-94: Improper Control of Generation of Code ('Code Inject…
Vendor / Product
vllm-project / vllm
Sources
cve.org  ·  NVD

Severity & Metrics

7.5 HIGH CVSS 3.1
CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H
Affected products (1)
VendorProductPlatformVersions
vllm-project vllm < 0.22.0
Weakness (CWE)
CWESourceDescription
CWE-617 cna CWE-617: Reachable Assertion
CWE-94 cna CWE-94: Improper Control of Generation of Code ('Code Injection')
CVSS scores (1)
ScoreSeverityVersionSourceVector
7.5 HIGH 3.1 cna CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:U/C:H/I:H/A:H
References (3)
Back to overview