CVE Database · CVE-2026-44223
CVSS v3.1
6.5
EPSS
0.37%
Published
May 12, 2026
Modified
May 15, 2026
Public PoC / Exploit
All weaponized →No public PoC or exploit code indexed for this CVE.
Links to public security research (Exploit-DB, Nuclei, Trickest, GitHub) for defensive use only.
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.
CVSS Vector
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:HWeaknesses (CWE)
Affected Products (1)
References (4)