CVE-2025-62372: Vllm Denial of Service Vulnerability

HIGH
Advisory Identifier: GHSA-pmqf-x6x8-p7qw
Share
CVSS 6.5
CWE-129
Published
Advisory Summary

CVE-2025-62372 is a HIGH (CVSS 6.5) denial of service vulnerability (CWE-129: Improper Validation of Array Index) affecting Vllm libraries in the PyPI ecosystem. The issue involves vllm vulnerable to dos with incorrect shape of multimodal embedding inputs. Remediation is available by upgrading affected packages to version 0.11.1.

CVE-2025-62372 at a Glance

CVE Identifier CVE-2025-62372
GHSA Identifier GHSA-pmqf-x6x8-p7qw
Severity HIGH
CVSS v3.1 Base Score 6.5
CWE Classification CWE-129: Improper Validation of Array Index
Affected Product Vllm
Package Ecosystem PyPI
Published Date
Last Updated

Security Impact

CVSS: 6.5

The security impact metrics below are parsed directly from the CVSS v3.1 vector string (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H).

Attack Vector Network (AV:N)
Attack Complexity Low (AC:L)
Privileges Required Low (PR:L)
User Interaction None (UI:N)
Scope Unchanged (S:U)
Confidentiality None (C:N)
Integrity None (I:N)
Availability High (A:H)

How to Fix CVE-2025-62372

To remediate this vulnerability, upgrade all affected Vllm dependencies to a patched release. Both direct and transitive dependencies should be checked, as frameworks or security libraries may pull in vulnerable versions transitively.

vllm fixed: 0.11.1 
pip install vllm>=0.11.1
Verify Direct and Transitive Dependencies

Run the following command in your project directory to verify that all resolved dependencies use the patched release: 

pip show vllm

How to Check if You Are Affected

To determine whether your Python environment is affected by CVE-2025-62372, inspect your installed packages using pip:

pip show vllm

Analyze the terminal output to locate any version of Vllm in your dependency tree. If any resolved version matches the vulnerable ranges listed in the table below (for example, versions earlier than 0.11.1), your application is affected.

Note: The pip show vllm command checks currently installed packages in your active Python environment.

Technical Details

Vulnerability Mechanism & Technical Analysis

This vulnerability is classified under CWE-129: Improper Validation of Array Index, where the product uses untrusted input when calculating or using an array index, but the product does not validate or incorrectly validates the index to ensure the index references a valid position within the array. In Vllm, software state validation or input boundaries are circumvented during processing. Review the official source advisory statement below for specific affected components, classes, and reproduction details.

Source Advisory Statement

### Summary Users can crash the vLLM engine serving multimodal models by passing multimodal embedding inputs with correct `ndim` but incorrect `shape` (e.g. hidden dimension is wrong), regardless of whether the model is intended to support such inputs (as defined in the Supported Models page). The issue has existed ever since we added support for image embedding inputs, i.e. #6613 (released in v0.5.5) ### Details Using image embeddings as an example: - For models that support image embedding inputs, the engine crashes when scattering the embeddings to `inputs_embeds` (mismatched shape) - For models that don't support image embedding inputs, the engine crashes when validating the inputs inside `get_input_embeddings` (validation fails). This happens because we only validate `ndim` of the tensor, but not the full shape, in input processor (via `MultiModalDataParser`). ### Impact - Denial of service by crashing the engine ### Mitigation - Use API key to limit access to trusted users. - Set `--limit-mm-per-prompt` to 0 for all non-text modalities to ban multimodal inputs, which includes multimodal embedding inputs. However, the model would then only accept text, defeating the purpose of using a multi-modal model. ### Resolution - https://github.com/vllm-project/vllm/pull/27204

CVE-2025-62372 Affected Vllm Versions

1 component(s)
Fixed Versions:
0.11.1
Product / Package Ecosystem Affected Versions Fixed Version
vllm
pypi/vllm
PyPI
≥ 0.5.5
< 0.11.1
0.11.1

References

GitHub Advisory / Wiki ADVISORY
https://github.com/vllm-project/vllm/security/advisories/GHSA-pmqf-x6x8-p7qw
NVD — CVE-2025-62372 NVD
https://nvd.nist.gov/vuln/detail/CVE-2025-62372
Vllm Source Repository REPOSITORY
https://github.com/vllm-project/vllm/pull/27204
Vllm Source Repository REPOSITORY
https://github.com/vllm-project/vllm/pull/6613
Security Fix Commit FIX COMMIT
https://github.com/vllm-project/vllm/commit/58fab50d82838d5014f4a14d991fdb9352c9c84b
GitHub Advisory / Wiki ADVISORY
https://github.com/advisories/GHSA-pmqf-x6x8-p7qw
Vllm Source Repository REPOSITORY
https://github.com/pypa/advisory-database/tree/main/vulns/vllm/PYSEC-2026-2019.yaml
Vllm Source Repository REPOSITORY
https://github.com/vllm-project/vllm
pypi.org/project/vllm WEB
https://pypi.org/project/vllm

Frequently Asked Questions

What is CVE-2025-62372?
CVE-2025-62372 is a security vulnerability in Vllm involving vLLM vulnerable to DoS with incorrect shape of multimodal embedding inputs under CWE-129: Improper Validation of Array Index.
Which Vllm versions are affected?
Affected versions include releases prior to 0.11.1 across affected packages: vllm.
What version fixes CVE-2025-62372?
Upgrading to 0.11.1 (or the latest patched release for your respective branch) resolves the vulnerability.
Is vllm affected?
Yes, vllm is affected in releases prior to 0.11.1. Upgrade your dependency to 0.11.1 or later.
What is the CVSS score of CVE-2025-62372?
CVE-2025-62372 has a CVSS v3.1 base score of 6.5 (HIGH) with vector string CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H.
What CWE is associated with CVE-2025-62372?
CVE-2025-62372 is classified under CWE-129: Improper Validation of Array Index (The product uses untrusted input when calculating or using an array index, but the product does not validate or incorrectly validates the index to ensure the index references a valid position within the array.).
How can I check PyPI dependencies for CVE-2025-62372?
Run the dependency tree command: 'pip show vllm'. This command lists direct and transitive dependencies in your build configuration so you can verify if a vulnerable version is resolved.
Machine-readable OSV JSON OSV v1.5 Schema
Normalized schema output
View Raw JSON 
{
  "schema_version" : "1.4.0",
  "id" : "GHSA-pmqf-x6x8-p7qw",
  "modified" : "2026-07-17T16:15:51Z",
  "published" : "2025-11-20T21:23:29Z",
  "aliases" : [ "CVE-2025-62372" ],
  "summary" : "vLLM vulnerable to DoS with incorrect shape of multimodal embedding inputs",
  "details" : "### Summary\n\nUsers can crash the vLLM engine serving multimodal models by passing multimodal embedding inputs with correct `ndim` but incorrect `shape` (e.g. hidden dimension is wrong), regardless of whether the model is intended to support such inputs (as defined in the Supported Models page).\n\nThe issue has existed ever since we added support for image embedding inputs, i.e. #6613 (released in v0.5.5)\n\n### Details\n\nUsing image embeddings as an example:\n\n- For models that support image embedding inputs, the engine crashes when scattering the embeddings to `inputs_embeds` (mismatched shape)\n- For models that don't support image embedding inputs, the engine crashes when validating the inputs inside `get_input_embeddings` (validation fails).\n\nThis happens because we only validate `ndim` of the tensor, but not the full shape, in input processor (via `MultiModalDataParser`).\n\n### Impact\n\n- Denial of service by crashing the engine\n\n### Mitigation\n\n- Use API key to limit access to trusted users.\n- Set `--limit-mm-per-prompt` to 0 for all non-text modalities to ban multimodal inputs, which includes multimodal embedding inputs. However, the model would then only accept text, defeating the purpose of using a multi-modal model.\n\n### Resolution\n\n- https://github.com/vllm-project/vllm/pull/27204",
  "severity" : [ {
    "type" : "CVSS_V3",
    "score" : "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H"
  }, {
    "type" : "CVSS_V4",
    "score" : "CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:H"
  } ],
  "affected" : [ {
    "package" : {
      "ecosystem" : "PyPI",
      "name" : "vllm"
    },
    "ranges" : [ {
      "type" : "ECOSYSTEM",
      "events" : [ {
        "introduced" : "0.5.5"
      }, {
        "fixed" : "0.11.1"
      } ]
    } ]
  } ],
  "references" : [ {
    "type" : "WEB",
    "url" : "https://github.com/vllm-project/vllm/security/advisories/GHSA-pmqf-x6x8-p7qw"
  }, {
    "type" : "ADVISORY",
    "url" : "https://nvd.nist.gov/vuln/detail/CVE-2025-62372"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/vllm-project/vllm/pull/27204"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/vllm-project/vllm/pull/6613"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/vllm-project/vllm/commit/58fab50d82838d5014f4a14d991fdb9352c9c84b"
  }, {
    "type" : "ADVISORY",
    "url" : "https://github.com/advisories/GHSA-pmqf-x6x8-p7qw"
  }, {
    "type" : "WEB",
    "url" : "https://github.com/pypa/advisory-database/tree/main/vulns/vllm/PYSEC-2026-2019.yaml"
  }, {
    "type" : "PACKAGE",
    "url" : "https://github.com/vllm-project/vllm"
  }, {
    "type" : "WEB",
    "url" : "https://pypi.org/project/vllm"
  } ],
  "database_specific" : {
    "cwe_ids" : [ "CWE-129" ],
    "severity" : "HIGH",
    "github_reviewed" : true,
    "github_reviewed_at" : "2025-11-20T21:23:29Z",
    "nvd_published_at" : "2025-11-21T02:15:43Z"
  }
}