CVE-2026-44223: 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…

PriorityP336medium6.5CVSS 3.1

AVNACLPRLUINSUCNINAH

EPSS

0.37%

28.5th percentile

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.

Affected

3 ranges

Vendor	Product	Version range	Fixed in
vllm-project	vllm	—	—
vllm	vllm	>= 0.18.0 < 0.20.0	0.20.0
vllm	vllm	>= 0.18.0 < 0.20.0	0.20.0

Stop checking back — get the weekly exploitation signal.

Every Monday: what got weaponized or added to CISA KEV in the last seven days — each CVE cross-linked to its PoC, Nuclei template, and detection rule. Free, one email a week, unsubscribe in one click.