Skip to content
LLMHigh

CVE-2026-22807: vLLM is an inference and serving engine for large language models (LLMs). Starting in version 0.10.1…

Occurred
Jan 21, 2026
Published
Jan 21, 2026
Source
NVD CVE

Summary

vLLM is an inference and serving engine for large language models (LLMs). Starting in version 0.10.1 and prior to version 0.14.0, vLLM loads Hugging Face `auto_map` dynamic modules during model resolution without gating on `trust_remote_code`, allowing attacker-controlled Python code in a model repo/path to execute at server startup. An attacker who can influence the model repo/path (local directory or remote Hugging Face repo) can achieve arbitrary code execution on the vLLM host during model load. This happens before any request handling and does not require API access. Version 0.14.0 fixes the issue.

Tags

  • CVE
  • CVE-2026-22807
  • cvss:HIGH

Classification

Category confidence: 95%
Severity confidence: 90%

Rationale: vLLM LLM serving engine vulnerability allows arbitrary code execution via untrusted model repos, representing a significant security risk.

Classifier: Qwen/Qwen3.5-35B-A3B-FP8

Visit source