CVE-2026-44222

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.
Configurations

Configuration 1 (hide)

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

History

14 May 2026, 15:38

Type Values Removed Values Added
CPE cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*
First Time Vllm vllm
Vllm
References () https://github.com/vllm-project/vllm/issues/32656 - () https://github.com/vllm-project/vllm/issues/32656 - Issue Tracking
References () https://github.com/vllm-project/vllm/security/advisories/GHSA-hpv8-x276-m59f - () https://github.com/vllm-project/vllm/security/advisories/GHSA-hpv8-x276-m59f - Exploit, Vendor Advisory

12 May 2026, 20:16

Type Values Removed Values Added
New CVE

Information

Published : 2026-05-12 20:16

Updated : 2026-05-14 15:38


NVD link : CVE-2026-44222

Mitre link : CVE-2026-44222

CVE.ORG link : CVE-2026-44222


JSON object : View

Products Affected

vllm

  • vllm
CWE
CWE-129

Improper Validation of Array Index