CVE-2026-126399: vLLM 0.19.0 before 0.31.0 Resource Exhaustion via Request-Level num_frames

Published Nov 13, 2026
·
Updated

vLLM from 0.19.0 before 0.31.0 contains a resource exhaustion vulnerability that allows inference clients to bypass the server frame limit by setting mediaiokwargs.video.numframes in chat completion requests. Attackers can supply numframes=-1 or a large value with many base64 JPEG frames, forcing VideoMediaIO to decode all frames and np.stack them, causing memory exhaustion.

Affected Software

1 affected component
vllm vllm>=0.19.0<0.31.0

Event History

Nov 13, 2026
CVE Published
via MITRE·09:35 AM
Data Sourced
via MITRE·09:35 AM
DescriptionSeverityWeakness
Data Sourced
via NVD·10:17 AM
DescriptionSeverityWeakness

Frequently Asked Questions

1

Who can exploit this issue?

An inference client with permission to submit chat completion requests can exploit it. The attack is network-reachable and requires low privileges, with no user interaction.

2

What request conditions trigger memory exhaustion?

The client supplies media_io_kwargs.video.num_frames as -1 or a large value and includes many base64-encoded JPEG frames. VideoMediaIO then decodes all supplied frames and stacks them in memory.

3

Are affected deployments protected by the server frame limit?

No. The request-level media_io_kwargs.video.num_frames setting can bypass the server frame limit in affected versions.

4

Which versions are affected?

vLLM versions from 0.19.0 before 0.31.0 are affected. Version 0.31.0 is not included in the affected range.

Contact

SecAlerts Pty Ltd.
132 Wickham Terrace
Fortitude Valley,
QLD 4006, Australia
info@secalerts.co
By using SecAlerts services, you agree to our services end-user license agreement. This website is safeguarded by reCAPTCHA and governed by the Google Privacy Policy and Terms of Service. All names, logos, and brands of products are owned by their respective owners, and any usage of these names, logos, and brands for identification purposes only does not imply endorsement. If you possess any content that requires removal, please get in touch with us.
© 2026 SecAlerts Pty Ltd.
ABN: 70 645 966 203, ACN: 645 966 203