CVE-2026-105758

Received Received - Intake

Memory Exhaustion in vLLM Video Processing

Publication date: 2026-10-05

Last updated on: 2026-10-05

Assigner: GitHub, Inc.

Description

vLLM is an inference and serving engine for large language models. From 0.24.0 until 0.30.0, the Qwen2VLVideoBackend and Qwen3VLVideoBackend classes accept request-level values for the media_io_kwargs.video.max_frames and media_io_kwargs.video.fps fields without enforcing server-side ceilings. An unauthenticated caller can submit these values to the /tokenize endpoint, causing the sampler to decode every frame selected from attacker-controlled video input, consume disproportionate frontend memory, and potentially terminate the API process before scheduling or admission control. The Rust frontend is not affected because it rejects the media_io_kwargs field. This issue is fixed in version 0.30.0.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published

2026-10-05

Last Modified

2026-10-05

Generated

2026-10-06

AI Q&A

2026-10-06

EPSS Evaluated

N/A

NVD

CVE-2026-105758

Affected Vendors & Products

Vendor	Product	Version / Range
vllm-project	vllm	>= 0.24.0, < 0.30.0

Helpful Resources

Exploitability

CWE

KEV

CWE ID	Description
CWE-770	The product allocates a reusable resource or group of resources on behalf of an actor without imposing any intended restrictions on the size or number of resources that can be allocated.

Attack-Flow Graph

Executive Summary

This vulnerability in vLLM versions 0.24.0 to 0.30.0 allows unauthenticated users to submit requests with high media_io_kwargs.video.max_frames and media_io_kwargs.video.fps values to the /tokenize endpoint. This causes excessive frame decoding from attacker-controlled video input, leading to high memory consumption and potential API process termination.

Detection Guidance

This vulnerability can be detected by checking the vLLM version in use. If your system runs versions between 0.24.0 and 0.30.0, it is potentially vulnerable. Inspect the installed package with commands like pip show vllm or checking version files in the deployment directory.

Impact Analysis

An attacker could exploit this to consume excessive server memory, causing service disruption or crashes. This may lead to downtime, degraded performance, or denial of service for legitimate users accessing the vLLM API.

Compliance Impact

This vulnerability could potentially impact compliance with GDPR and HIPAA by enabling denial-of-service attacks that disrupt service availability. Uncontrolled resource consumption may lead to service unavailability, which could violate availability requirements in GDPR Article 32 and HIPAA Security Rule Section 164.308(a)(7).

Mitigation Strategies

Upgrade vLLM to version 0.30.0 or later immediately. If upgrading is not possible, restrict access to the /tokenize endpoint to prevent unauthenticated calls. Monitor API processes for unexpected memory consumption or crashes.

Hi! I’m here to help you understand CVE-2026-105758. Ask me anything about the vulnerability, its impact, or mitigation strategies.

0/70