CVEvulnerability

CVE-2026-44222

7.5 HIGH Published 2026-05-12 Modified 2026-06-17 Official page ↗
Status
Analyzed
Published
2026-05-12
Last modified
2026-06-17
Assigner
[email protected]
CVSS 3.1 ([email protected])
6.5 MEDIUM — CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
CVSS 3.1 ([email protected])
7.5 HIGH — CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
Weaknesses
CWE-129

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.

Affected configurations

• vllm vllm (>= 0.6.1 < 0.20.0)

Related

Weakness · CWE (1)

  • CWE-129 Improper Validation of Array Index weakness

References