CVE-2026-126399

Description

A flaw was found in vLLM. A remote attacker can trigger a Denial of Service (DoS) by sending a specially crafted chat completion request with an unrestricted video frame count alongside numerous video frames. Because request-level parameters override the server-side frame limit, the service attempts to decode and process all submitted frames, leading to excessive memory consumption and potential service failure.

Statement

This vulnerability is rated as having Moderate impact because it results solely in a denial of service through memory exhaustion without compromising system confidentiality or integrity. Exploitation is limited to environments hosting multimodal models configured to process video inputs, meaning deployments dedicated exclusively to text-only inference remain unaffected. While an authorized client submitting chat completion queries can cause the inference server process to terminate due to excessive memory consumption, the flaw cannot be leveraged for code execution or unauthorized data access.

Mitigation

Restrict network reachability to the inference server so that only trusted clients and services can access the API endpoint. In environments using firewalld, access to the vLLM service port (default TCP port 8000) can be restricted to specific trusted source IP addresses:

# firewall-cmd --zone=trusted --add-source=<trusted_ip_or_subnet> --permanent
# firewall-cmd --zone=public --remove-port=8000/tcp --permanent
# firewall-cmd --reload

Alternatively, deploy an API gateway or reverse proxy in front of the vLLM server to enforce client authentication and inspect incoming requests to block untrusted media processing parameters.

Caveats:
Restricting network access limits API availability strictly to designated clients. Modifying and reloading firewall rules may temporarily disrupt active network connections.

Common Vulnerability Scoring System (CVSS) Score Details

Info alert:Important note

CVSS scores for open source components depend on vendor-specific factors (e.g. version or build chain). Therefore, Red Hat's score and impact rating can be different from NVD and other vendors. Red Hat remains the authoritative CVE Naming Authority (CNA) source for its products and services (see Red Hat classifications).

The following CVSS metrics and score provided are preliminary and subject to review.

CVSS v3 Score Breakdown

Red HatNVDcve.org
Base Score6.5N/A6.5
Attack VectorNetworkN/ANetwork
Attack ComplexityLowN/ALow
Privileges RequiredLowN/ALow
User InteractionNoneN/ANone
ScopeUnchangedN/AUnchanged
ConfidentialityNoneN/ANone
Integrity ImpactNoneN/ANone
Availability ImpactHighN/AHigh

Vector

Red Hat: CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H

cve.org: CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H

Understanding the Weakness (CWE)

Availability

Technical Impact: DoS: Resource Consumption (CPU); DoS: Resource Consumption (Memory); DoS: Resource Consumption (Other)

When allocating resources without limits, an attacker could prevent other systems, applications, or processes from accessing the same type of resource. It can be easy for an attacker to consume many resources by rapidly making many requests or causing larger resources to be used than is needed.

Frequently Asked Questions

Want to get errata notifications? Sign up here.