---
id: CVE-2026-56340
title: >-
  vLLM versions >= 0.10.2 and < 0.13.0 are missing sparse tensor validation in
  multimodal embeddings processing
summary: >-
  vLLM versions >= 0.10.2 and < 0.13.0 are missing sparse tensor validation in
  multimodal embeddings processing. Because PyTorch disables sparse tensor
  invariant checks by default, an attacker can submit crafted embedding requests
  with mal…
severity: high
cvss: 8.8
cvssVector: 'CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H'
cwe:
  - CWE-20
  - CWE-787
vendor: vllm
product: vllm
affected:
  - 'vllm >= 0.10.2, < 0.13.0'
patched:
  - vllm 0.13.0
published: '2026-06-20'
updated: '2026-07-15'
sourceUpdated: '2026-07-15T01:16:38.343'
source: NVD
sourceUrl: 'https://nvd.nist.gov/vuln/detail/CVE-2026-56340'
references:
  - url: >-
      https://github.com/vllm-project/vllm/security/advisories/GHSA-mcmc-2m55-j8jj
    label: disclosure@vulncheck.com
  - url: >-
      https://www.vulncheck.com/advisories/vllm-denial-of-service-via-unvalidated-multimodal-embeddings
    label: disclosure@vulncheck.com
  - url: 'https://access.redhat.com/security/cve/CVE-2026-56340'
    label: 0b0ca135-0b70-47e7-9f44-1890c2a1c46c
  - url: 'https://bugzilla.redhat.com/show_bug.cgi?id=2491060'
    label: 0b0ca135-0b70-47e7-9f44-1890c2a1c46c
  - url: >-
      https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-56340.json
    label: 0b0ca135-0b70-47e7-9f44-1890c2a1c46c
tags:
  - nvd
  - cve.org
ssvc:
  exploitation: none
  automatable: 'no'
  technicalImpact: total
  timestamp: '2026-06-24T15:54:59.666959Z'
ingestedAt: '2026-09-14T00:23:31.573Z'
epss: 0.00644
epssPercentile: 0.48606
---

## Overview

vLLM versions >= 0.10.2 and < 0.13.0 are missing sparse tensor validation in multimodal embeddings processing. Because PyTorch disables sparse tensor invariant checks by default, an attacker can submit crafted embedding requests with malformed (negative or out-of-bounds) tensor indices, when the prompt-embeds feature is enabled, to trigger crashes or resource exhaustion (denial of service), with potential for out-of-bounds/write-what-where memory corruption. This continues CVE-2025-62164, whose prior fix only disabled the feature by default rather than addressing the root cause.

## Affected

- `vllm >= 0.10.2, < 0.13.0`

## Remediation

Upgrade past the affected range:

- `vllm 0.13.0`