vLLM Vulnerabilities

pipLLM Inference

AI Threat Alert tracks 188 known vulnerabilities in vLLM, 13 rated critical — an AI/ML llm inference in the pip ecosystem. Each CVE includes CVSS severity, EPSS exploit probability, patch status, and CISO-grade analysis.

61
Risk Score
188
Total CVEs
13
Critical
pip
Ecosystem
Nov 18, 2026
Last CVE
26%
Patch Rate
37d
Avg Time to Patch
96,942 stars24,834 forks10,395 issues93 dependentsLast push Nov 15, 2026
View on GitHub

Known Vulnerabilities(188 total, page 2 of 8)

SeverityCVE IDSummaryCVSSPublished
MEDIUMCVE-2026-116175vLLM: audio decompression bomb exhausts server memory6.5Oct 25, 2026HIGHCVE-2026-116174vLLM: gRPC invalid token IDs crash EngineCore (DoS)7.5Oct 25, 2026HIGHCVE-2026-116173vLLM: gRPC logprob limit bypass crashes EngineCore7.5Oct 25, 2026HIGHCVE-2026-116172vLLM: gRPC top_n validation flaw crashes EngineCore7.5Oct 25, 2026HIGHCVE-2026-114055vLLM: unvalidated mm_processor_kwargs enables remote DoS7.5Oct 21, 2026MEDIUMCVE-2026-114054vLLM: cache aliasing leaks cross-request media/KV state6.5Oct 21, 2026HIGHCVE-2026-114053vLLM: RCE via code_revision in /tokenize endpoint8.1Oct 21, 2026HIGHCVE-2026-114052vLLM: unauthenticated /tokenize exhausts memory (MOSS-Audio)7.5Oct 21, 2026HIGHCVE-2026-114051vLLM: unauth DoS via min_tokens hangs generation7.5Oct 21, 2026HIGHCVE-2026-114050vLLM: forged multimodal UUID crashes inference engine7.5Oct 21, 2026LOWCVE-2026-114049vLLM: stale token mask leaks across requests3.7Oct 21, 2026CRITICALCVE-2026-105192LMCache: unauthenticated RCE via pickle in ZMQ transport9.8Oct 16, 2026HIGHCVE-2026-14164libarchive: RAR5 double free enables remote DoS7.5Jun 30, 2026MEDIUMCVE-2026-109311vLLM: heap overflow DoS via IAMF audio upload6.5Oct 12, 2026MEDIUMCVE-2026-105922vLLM: penalty handler flaw allows remote DoS4.3Oct 6, 2026MEDIUMCVE-2026-105775vLLM: OOB read in Mamba2 mixer enables remote DoS4.3Oct 6, 2026MEDIUMCVE-2026-105756vLLM: unvalidated cache_salt crashes EngineCore (DoS)6.5Oct 5, 2026MEDIUMCVE-2026-105755vLLM: request-ID cache key collision corrupts rerank scores4.2Oct 5, 2026MEDIUMCVE-2026-105757vLLM: unhandled structured-output errors crash engine (DoS)6.5Oct 5, 2026MEDIUMCVE-2026-105754vLLM: input validation flaws let attacker crash EngineCore6.5Oct 5, 2026MEDIUMCVE-2026-105760vLLM: unbounded video frame options enable DoS5.3Oct 5, 2026MEDIUMCVE-2026-105758vLLM: unbounded video frame decoding enables DoS5.3Oct 5, 2026HIGHCVE-2026-105759vLLM: unbounded metrics labels allow memory-exhaustion DoS7.5Oct 5, 2026MEDIUMCVE-2026-105753vLLM: reachable assertion in cache causes DoS6.5Oct 5, 2026LOWCVE-2026-105752vLLM: cache_salt dropped on tool continuations leaks prefixes3.1Oct 5, 2026

Showing 26–50 of 188

Frequently asked questions

What is vLLM?

vLLM is an AI/ML llm inference tracked by AI Threat Alert for security vulnerabilities in the pip ecosystem.

How many known vulnerabilities does vLLM have?

vLLM has 188 known CVEs, 13 of them critical, tracked from NVD and GitHub Advisory.

Which ecosystem is vLLM distributed in?

vLLM is distributed via the pip ecosystem and categorized as llm inference.

Where does the vLLM vulnerability data come from?

Vulnerability data is sourced from NVD and GitHub Advisory, enriched with CVSS, EPSS, exploit signals, and patch status for each CVE.

How do I assess the risk of vLLM?

Review each CVE below — every entry shows CVSS severity, EPSS exploit probability, exploitation signals, and whether a patched version is available.

Monitor vLLM in your stack

Get instant alerts when new vulnerabilities affect vLLM. CISO analysis, ATLAS technique mappings, and compliance reports included.

Start Monitoring