CVE-2026-34753
Published:April 06, 2026
Updated:April 20, 2026
vLLM is an inference and serving engine for large language models (LLMs). From 0.16.0 to before 0.19.0, a server-side request forgery (SSRF) vulnerability in download_bytes_from_url allows any actor who can control batch input JSON to make the vLLM batch runner issue arbitrary HTTP/HTTPS requests from the server, without any URL validation or domain restrictions.
This can be used to target internal services (e.g. cloud metadata endpoints or internal HTTP APIs) reachable from the vLLM host. This vulnerability is fixed in 0.19.0.
Affected Packages
vllm (CONDA):
Affected version(s) >=0.8.3 <0.19.0Fix Suggestion:
Update to version 0.19.0https://github.com/vllm-project/vllm.git (GITHUB):
Affected version(s) >=v0.1.0 <v0.18.1Fix Suggestion:
Update to version v0.18.1vllm (PYTHON):
Affected version(s) >=0.16.0 <0.19.0Fix Suggestion:
Update to version 0.19.0Related Resources (5)
Do you need more information?
Contact UsCVSS v4
Base Score:
5.3
Attack Vector
NETWORK
Attack Complexity
LOW
Attack Requirements
NONE
Privileges Required
LOW
User Interaction
NONE
Vulnerable System Confidentiality
LOW
Vulnerable System Integrity
NONE
Vulnerable System Availability
LOW
Subsequent System Confidentiality
NONE
Subsequent System Integrity
NONE
Subsequent System Availability
NONE
CVSS v3
Base Score:
5.4
Attack Vector
NETWORK
Attack Complexity
LOW
Privileges Required
LOW
User Interaction
NONE
Scope
UNCHANGED
Confidentiality
LOW
Integrity
NONE
Availability
LOW
Weakness Type (CWE)
Server-Side Request Forgery (SSRF)
EPSS
Base Score:
0.04
