vllm Known Vulnerabilities & Fixes
vllm has 6 known security vulnerabilities in PyPI (Python). Upgrade to version 0.14.1 or later to resolve all known issues. Data sourced from OSV, enriched with EPSS exploit probability and CISA KEV.
Low Immediate Risk
No actively exploited vulnerabilities detected. Monitor and update in your next maintenance window.
Recommended safe version: 0.14.1
Upgrading to 0.14.1 or later resolves all 6 known vulnerabilities in vllm. Run: pip install "vllm>=0.14.1"
Is vllm in your project?
Check if you're affected and upgrade to 0.14.1 to stay secure.
Security Summary
- 6 affected version ranges have been identified across these 6 advisories.
- 50% of known issues have a published fixed version.
- Remediation: pip install --upgrade vllm
Vulnerabilities
6 unique vulnerabilities — sorted by exploitation risk (KEV → EPSS → CVSS). Click a CVE/GHSA ID for full details. First disclosed Mar 20, 2025, most recently Feb 2, 2026 (319 days apart).
| CVE / GHSA | Severity | Affected | Fixed In |
|---|---|---|---|
| CVE-2025-47277 vLLM Allows Remote Code Execution via PyNcclPipe Communication Service | CRITICAL | 0.6.5, 0.6.6, 0.6.6.post1, 0.7.0 (+8 more) | 0.8.5 |
| CVE-2024-9053 vLLM allows Remote Code Execution by Pickle Deserialization via AsyncEngineRPCServer() RPC server entrypoints | CRITICAL | 0.0.1, 0.1.0, 0.1.1, 0.1.2 (+32 more) | No fix available |
| CVE-2024-9052 vLLM deserialization vulnerability in vllm.distributed.GroupCoordinator.recv_object | CRITICAL | 0.0.1, 0.1.0, 0.1.1, 0.1.2 (+49 more) | No fix available |
| CVE-2024-11041 vLLM Deserialization of Untrusted Data vulnerability | CRITICAL | 0.0.1, 0.1.0, 0.1.1, 0.1.2 (+36 more) | No fix available |
| CVE-2026-22778 vLLM has RCE In Video Processing | CRITICAL | 0.10.0, 0.10.1, 0.10.1.1, 0.10.2 (+14 more) | 0.14.1 |
| CVE-2025-32444 vLLM Vulnerable to Remote Code Execution via Mooncake Integration | CRITICAL | 0.6.5, 0.6.6, 0.6.6.post1, 0.7.0 (+8 more) | 0.8.5 |
About This Data
Vulnerability data for vllm is sourced from the Open Source Vulnerability (OSV) database, aggregating reports from GitHub Advisory Database, NIST NVD, and ecosystem-specific sources.
CVSS (Common Vulnerability Scoring System) scores reflect exploitability and impact. EPSS (Exploit Prediction Scoring System) scores indicate the probability of exploitation within the next 30 days. Vulnerabilities marked with are listed in the CISA Known Exploited Vulnerabilities (KEV) catalog.
Related PyPI (Python) Packages
Other packages in this ecosystem, ranked by shared vulnerabilities where available.
Check Your Dependencies
Scan your project to check if you're using a vulnerable version of vllm.