CVE-2026-44222

HIGH EPSS 33.1%
Published May 12, 20261mo ago · Modified Jun 17, 20261w ago
7.5 CVSS 3.1
High
Find Similar
Published May 12, 2026 1mo ago
Last Modified Jun 17, 2026 1w ago

Description

vLLM is an inference and serving engine for large language models (LLMs). From 0.6.1 to before 0.20.0, there is a a Token Injection vulnerability in vLLM’s multimodal processing. Unauthenticated, text-only prompts that spell special tokens are interpreted as control. Image and video placeholder sequences supplied without matching data cause vLLM to index into empty grids during input-position computation, raising an unhandled IndexError and terminating the worker or degrading availability. Multimodal paths that rely on image_grid_thw/video_grid_thw are affected. This vulnerability is fixed in 0.20.0.

CVSS Details

Base Score
7.5
Exploitability
3.9
Impact
3.6
Vector string
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
Attack Vector Network
Attack Complexity Low
Privileges Required None
User Interaction None
Scope Unchanged
Confidentiality None
Integrity None
Availability High

Threat Intelligence

EPSS Exploit Probability
33.1% percentile
Exploit & Patch Status
Public Exploit Known
No Patch Available

Weaknesses 1

CWE-129

Affected Products 1

VendorProductVersionRange
vllmvllm*≥0.6.1  –  <0.20.0

References 2

  • github.com https://github.com/vllm-project/vllm/issues/32656
    Issue Tracking
  • github.com https://github.com/vllm-project/vllm/security/advisories/GHSA-hpv8-x276-m59f
    ExploitVendor Advisory

Remediation

No remediation data recorded yet

Check vendor advisories and the NVD entry for patch availability.