CVE-2026-5497: Unbounded Frame Count in video/jpeg Base64 Data URL Processing Leads to OOM DoS in vllm-project/vllm

Published Jun 11, 2026
·
Updated

vLLM versions 0.8.0 and later are vulnerable to an Out-of-Memory (OOM) Denial of Service (DoS) attack due to unbounded frame count processing in the VideoMediaIO.loadbase64() method. When processing video/jpeg data URLs, the method splits the base64 data string on commas to extract individual JPEG frames without enforcing a frame count limit. An attacker can exploit this by crafting a single API request containing thousands of comma-separated base64-encoded JPEG frames in a data URL, causing the server to decode all frames into memory and crash due to excessive memory consumption. This vulnerability is reachable via the OpenAI-compatible chat completions API and does not require authentication.

Affected Software

2 affected components
pypi/vllm>=0.8.0
vllm vllm>=0.8.0<0.19.0

Remediation

Recommended actions to resolve this vulnerability, in priority order.

  1. Configuration

    In vllm-project/vllm, modify the `VideoMediaIO.load_base64()` logic that processes `video/jpeg` data URLs to enforce a maximum frame count when splitting the base64 string on commas, instead of processing an unbounded number of frames (vulnerable in vLLM versions 0.8.0 and later).

    vllm-project/vllm video/jpeg data URL frame count limit (enforce maximum frames) = max_frames (set a strict upper bound)
  2. Compensating control

    Add an external rate limit for the OpenAI-compatible chat completions API endpoint to reduce the impact of unauthenticated abuse delivering extremely large video/jpeg data URLs with thousands of base64 frames.

Event History

Jun 11, 2026
CVE Published
via MITRE·08:31 AM
Data Sourced
via MITRE·08:31 AM
DescriptionSeverityWeakness
Data Sourced
via Red Hat·10:01 AM
DescriptionSeverityAffected Software
Data Sourced
via NVD·10:16 AM
RemedyDescriptionSeverityWeaknessAffected Software
Free Weekly Intel

Don't miss critical vulnerabilities

Join thousands of security professionals who receive our weekly digest of trending CVEs, zero-days, and exploited vulnerabilities.

No spam. Unsubscribe anytime.

Frequently Asked Questions

1

What is the severity of CVE-2026-5497?

CVE-2026-5497 has a severity rating of high with a score of 7.5.

2

How do I fix CVE-2026-5497?

To mitigate CVE-2026-5497, update to a patched version of vLLM that addresses the unbounded frame count in the VideoMediaIO.load_base64() method.

3

What kind of attack does CVE-2026-5497 facilitate?

CVE-2026-5497 can facilitate an Out-of-Memory (OOM) Denial of Service (DoS) attack.

4

Which versions of vLLM are affected by CVE-2026-5497?

vLLM versions 0.8.0 and later are affected by CVE-2026-5497.

5

What does the term 'unbounded frame count' mean in relation to CVE-2026-5497?

In the context of CVE-2026-5497, 'unbounded frame count' refers to the method's inability to limit the number of frames it processes, potentially leading to excessive memory usage.

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