Denial of Service Vulnerability in vLLM Inference Engine
CVE-2026-115911

6.5MEDIUM

Key Information:

Vendor

Vllm-project

Status
Vllm
Vendor
CVE Published:
26 October 2026

What is CVE-2026-115911?

The vLLM inference and serving engine for large language models is susceptible to a Denial of Service attack when no server-side chat template is configured. Authenticated users can supply a request chat template, or a compromised model repository can deliver a template that is processed without appropriate limits. If rendering times out or expands outputs excessively, it may block crucial endpoints, leading to performance degradation or memory exhaustion. This risk is mitigated in version 0.32.0.

Affected Version(s)

vllm >= 0.19.0, < 0.32.0

References

https://github.com/vllm-project/vllm/security/advisories/...
x_refsource_CONFIRM
https://github.com/vllm-project/vllm/pull/50300
x_refsource_MISC
https://github.com/vllm-project/vllm/commit/cff08b461e086...
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

.