Denial of Service Vulnerability in vLLM Large Language Model Engine
CVE-2026-73559

6.5MEDIUM

Key Information:

Vendor

Vllm-project

Status
Vllm
Vendor
CVE Published:
13 August 2026

What is CVE-2026-73559?

The vLLM inference and serving engine for large language models has a flaw that allows an authenticated API client to send crafted completion requests. This results in an unbounded number of processing elements being created, which can exhaust system resources like CPU, memory, and async scheduling capacity. The vulnerability was present in versions 0.19.0 through 0.26.0, and has been addressed in the latest version 0.26.0.

Affected Version(s)

vllm >= 0.19.0, < 0.26.0

References

https://github.com/vllm-project/vllm/security/advisories/...
x_refsource_CONFIRM
https://github.com/vllm-project/vllm/pull/47845
x_refsource_MISC
https://github.com/vllm-project/vllm/commit/675f4295cdfe0...
x_refsource_MISC

CVSS V3.1

Score:
6.5
Severity:
MEDIUM
Confidentiality:
None
Integrity:
None
Availability:
None
Attack Vector:
Network
Attack Complexity:
Low
Privileges Required:
Low
User Interaction:
None
Scope:
Unchanged

Timeline

  • Vulnerability published

  • Vulnerability Reserved

.