---
id: CVE-2026-100650
title: >-
  vLLM through 0.29.0 fetches and fully materializes remote or inline media
  before enforcing its documented media controls (the
  VLLM_MAX_AUDIO_CLIP_FILESIZE_MB compressed-audio size cap, default 25 MB, and
  the per-modality --limit-mm-per-p…
summary: >-
  vLLM through 0.29.0 fetches and fully materializes remote or inline media
  before enforcing its documented media controls (the
  VLLM_MAX_AUDIO_CLIP_FILESIZE_MB compressed-audio size cap, default 25 MB, and
  the per-modality --limit-mm-per-p…
severity: medium
cvss: 6.5
cvssVector: 'CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H'
cwe:
  - CWE-400
vendor: vllm-project
product: vllm
affected:
  - vllm < 0.29.0
published: '2026-09-26'
updated: '2026-09-26'
sourceUpdated: '2026-09-26T14:16:47.523'
source: NVD
sourceUrl: 'https://nvd.nist.gov/vuln/detail/CVE-2026-100650'
references:
  - url: >-
      https://github.com/vllm-project/vllm/commit/752a3a504485790a2e8491cacbb35c137339ad34
    label: disclosure@vulncheck.com
  - url: >-
      https://github.com/vllm-project/vllm/security/advisories/GHSA-p6g9-7v3x-m8mv
    label: disclosure@vulncheck.com
  - url: >-
      https://www.vulncheck.com/advisories/vllm-before-0.29.0-resource-exhaustion-via-unbounded-media-materialization
    label: disclosure@vulncheck.com
tags:
  - nvd
  - cve.org
ingestedAt: '2026-09-26T13:33:20.920Z'
---

## Overview

vLLM through 0.29.0 fetches and fully materializes remote or inline media before enforcing its documented media controls (the VLLM_MAX_AUDIO_CLIP_FILESIZE_MB compressed-audio size cap, default 25 MB, and the per-modality --limit-mm-per-prompt item limits). Across four ingress paths — the shared media-acquisition layer (HTTPConnection.get_bytes()/async_get_bytes()), the chat completions audio_url/base64 path, the batch speech runner, and the Rust frontend POST /tokenize route — the server reads the entire HTTP response body, base64-decodes the inline payload, or spawns one fetch/decode task per media part, and only then applies the limit (or, on some paths, never applies it). A remote attacker can therefore cause the API server or batch-runner process to allocate memory and consume outbound bandwidth proportional to an attacker-chosen body size or media item count before the request is rejected, resulting in pre-inference memory and bandwidth exhaustion (denial of service). The chat and batch surfaces require an API key when one is configured; the Rust frontend /tokenize route is unauthenticated by design. There is no code execution or data disclosure impact.

## Remediation

Refer to the linked advisories for vendor-supplied fixes and affected version ranges.