CVE-2026-116178

vLLM from 0.20.0 before 0.30.0 contains a denial of service vulnerability that allows remote attackers to crash the engine by sending inconsistent multimodal metadata to the disaggregated /inference/v1/generate endpoint. Attackers can replay a valid multimodal item with repeated or overlapping mm_placeholders ranges or mismatched kwargs_data and mm_hashes lists, producing a negative text span in Qwen2.5-VL position construction that terminates the engine.
Configurations

Configuration 1 (hide)

cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*

History

28 Oct 2026, 15:40

Type Values Removed Values Added
CPE cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*
References () https://github.com/vllm-project/vllm/security/advisories/GHSA-34qw-q2h2-jp3m - () https://github.com/vllm-project/vllm/security/advisories/GHSA-34qw-q2h2-jp3m - Patch, Vendor Advisory
References () https://www.vulncheck.com/advisories/vllm-0.20.0-before-0.30.0-dos-via-inference-v1-generate-mm-placeholders - () https://www.vulncheck.com/advisories/vllm-0.20.0-before-0.30.0-dos-via-inference-v1-generate-mm-placeholders - Third Party Advisory
First Time Vllm vllm
Vllm

25 Oct 2026, 14:16

Type Values Removed Values Added
New CVE

Information

Published : 2026-10-25 14:16

Updated : 2026-10-28 15:40


NVD link : CVE-2026-116178

Mitre link : CVE-2026-116178

CVE.ORG link : CVE-2026-116178


JSON object : View

Products Affected

vllm

  • vllm
CWE
CWE-20

Improper Input Validation