HIGHVulnerability
Verified
Global

NVD HIGH: CVE-2026-116178 — vLLM from 0.20.0 before 0.30.0 contains a denial of service vulnerability that a...

·Source: NIST NVD

Updated:

Executive Summary

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 Q

Analysis

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. CVSS Score: 5.9. Published: 2026-10-25T14:16:57.483.

Indicators of Compromise (1)

CVE (1)
CVE-2026-116178
NVDMITRE CVE
Source Attribution

Originally published by NIST NVD on Oct 25, 2026. Verified by: NIST.

Related Threats

CRITICALApt

Siemens Mendix SAML Module

<p><a href="https://github.com/cisagov/CSAF/blob/develop/csaf_files/OT/white/2026/icsa-26-302-01.json"><strong>View CSAF</strong></a></p> <h2>Summary</h2> <p><strong>Mendix SAML module contains an authentication bypass vulnerability that could allow an unauthenticated remote attacker to gain unauthorized access by replaying a valid SAML assertion in IdP-initiated authentication flows. Siemens has

CVE-2026-89209
CISA Advisories
CRITICALMalware

Deploying a Local Ghidra Server for Your Malware Analysis Lab, (Thu, Oct 29th)

Building your malware analysis lab is a key point before playing with malware. Today, thanks to virtualization, it&&#x23&#x3b;x26&#x3b;&#x23&#x3b;39&#x3b;s convenient to play with snapshots in your lab. Start a new investigation from a clean snapshot and take some more when you reached a critical step. But, if you need to revert to a previous snapshot, all changes applied to the VM are... lost&#x2

SANS ISC
MEDIUMVulnerability

LABScon26 Replay | Keynote: A Sober Analysis of AI Impact

Harvard Professor James Mickens weighs the promises of AI against its potential harms, arguing for hardware-based mechanisms to isolate frontier models.

SentinelLabs