CVE-2026-44223
CVE-2026-44223 · vLLM: extract_hidden_states speculative decoding crashes server on any request with penalty parameters
Cette page présente des faits issus de sources publiques. Consultez l’avis original avant toute mise à jour.
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 · Moyenne · EPSS 0,4% · Correctif disponible
- Publication
- Mise à jour
Pile logicielle: python