p PatchFirst

CVE-2026-44223

CVE-2026-44223 · vLLM: extract_hidden_states speculative decoding crashes server on any request with penalty parameters

This page contains facts from public vulnerability sources. Check the original advisory before updating.

vLLM is an inference and serving engine for large language models (LLMs). From 0.18.0 to before 0.20.0, the extract_hidden_states speculative decoding proposer in vLLM returns a tensor with an incorrect shape after the first decode step, causing a RuntimeError that crashes the EngineCore process. The crash is triggered when any request in the batch uses sampling penalty parameters (repetition_penalty, frequency_penalty, or presence_penalty). A single request with a penalty parameter (e.g., "repetition_penalty": 1.1) is sufficient to crash the server. This vulnerability is fixed in 0.20.0.

CVSS 6.5 · Medium · EPSS 0.4% · Fix available

Published
Updated

Stack: python

Affected products and versions

  • vllm

    Versions
    • >= 0.18.0, < 0.20.0: Affected

    No fix published

  • vllm PyPI

    Versions
    • 0.18.0
    • 0.18.1
    • 0.19.0
    • 0.19.1
    Affected ranges
    • ECOSYSTEM: ≥ 0.18.0, < 0.20.0

    Update to 0.20.0

Original sources

Check your server →