CVE catalog

CVE-2026-105758

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

[CVE catalog]

Severity
Medium · CVSS 5.3CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L
Published
Updated
Source
[NVD] · [CVE record]
EPSS
0.30% · Oct 25, 2026

A flaw was found in vllm. An unauthenticated remote attacker can cause a Denial of Service (DoS) by sending requests with excessively high video processing parameters to affected API endpoints. Because the backend does not enforce upper bounds on user-controlled video frame counts and rates, the service attempts to decode all frames from attacker-provided video data. This behavior consumes disproportionate server memory and can terminate the API process, making the service unavailable. vllm: vllm: Denial of Service via unbounded video frame parameters A MODERATE severity flaw was found in vLLM, as shipped in Red Hat AI Inference Server 3 and Red Hat OpenShift AI (RHOAI). 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 upper bounds. An unauthenticated remote attacker can submit requests with excessively high video frame counts or sampling rates to API endpoints such as /tokenize, causing the service to decode every frame selected from attacker-controlled video input. This consumes disproportionate frontend memory and can terminate the API process before scheduling or admission control, resulting in a denial of service with no impact on confidentiality or integrity (CVSS 5.3, AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L, matching the upstream rating). The Rust frontend is not affected because it rejects the media_io_kwargs field. For the affected Red Hat AI Inference Server and RHOAI products the fix is deferred because the impact is MODERATE with a CVSS score below 7.0; the residual risk is limited to availability, where an unauthenticated remote attacker could crash the vLLM API process and make the inference service unavailable until it is restarted. The CVSS score(s) listed for this vulnerability do not reflect the associated product's status, and are included for informational purposes to better understand the severity of this vulnerability.

Vulnerable and fixed packages

Distribution-reported package versions and fixes. These do not depend on matching against the OpenFactory package catalog.

DistributionPackageReleaseVersionStatus
CpevllmNot reported0.30.0fixed
PyPIvllmNot reported0.24.0vulnerable
PyPIvllmNot reported0.25.0vulnerable
PyPIvllmNot reported0.25.1vulnerable
PyPIvllmNot reported0.26.0vulnerable
PyPIvllmNot reported0.27.0vulnerable
PyPIvllmNot reported0.27.1vulnerable
PyPIvllmNot reported0.28.0vulnerable
PyPIvllmNot reported0.29.0vulnerable
PyPIvllmNot reported0.30.0fixed
PyPIvllmNot reported0.24.0vulnerable
PyPIvllmNot reported0.25.0vulnerable
PyPIvllmNot reported0.25.1vulnerable
PyPIvllmNot reported0.26.0vulnerable
PyPIvllmNot reported0.27.0vulnerable
PyPIvllmNot reported0.27.1vulnerable
PyPIvllmNot reported0.28.0vulnerable
PyPIvllmNot reported0.29.0vulnerable
PyPIvllmNot reported0.30.0fixed

Fixed-version evidence

DistributionPackageTypeReleaseFixed versionUrgencyOriginBugs
CpevllmcpeNot reported0.30.0Not reportedNVD-CVE-2026-105758Not reported
PyPIvllmPyPINot reported0.30.0Not reportedPYSEC-2026-4310Not reported
PyPIvllmPyPINot reported0.30.0Not reportedGHSA-X6MC-67GF-CHW4Not reported

OpenFactory catalog packages

No matching packages in this catalog.

Observed:

References