Home

Description

vLLM is an inference and serving engine for large language models (LLMs). From 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.

PUBLISHED Reserved 2026-05-05 | Published 2026-05-12 | Updated 2026-05-15 | Assigner GitHub_M




MEDIUM: 6.5CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H

Problem types

CWE-131: Incorrect Calculation of Buffer Size

CWE-704: Incorrect Type Conversion or Cast

Product status

>= 0.18.0, < 0.20.0
affected

References

github.com/...t/vllm/security/advisories/GHSA-83vm-p52w-f9pw exploit

github.com/vllm-project/vllm/pull/38610 exploit

github.com/...t/vllm/security/advisories/GHSA-83vm-p52w-f9pw

github.com/vllm-project/vllm/pull/38610

cve.org (CVE-2026-44223)

nvd.nist.gov (CVE-2026-44223)

Download JSON