CVE-2026-115914: vllm — Fix

Description

vLLM is an inference and serving engine for large language models. From 0.11.1 until 0.32.0, the DeepSeek-OCR, DeepSeek-OCR-2, and Unlimited-OCR image path allows a remote client that can reach the OpenAI-compatible API to submit a legal image whose normalized pixels sum to zero when that image is alone in its per-step encoder batch. _parse_and_validate_image_input returns no embedding, embed_multimodal propagates that value, and sanity_check_mm_encoder_outputs triggers a fatal assertion that shuts down the engine for all users. This issue is fixed in version 0.32.0.

Is your site exposed to CVE-2026-115914?

Run a free security scan — no signup, results in seconds.

CVSS v3.1 Score

6.5

MEDIUM

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

EPSS — Exploit Prediction

0.0040

Probability of exploitation

0.34%

Percentile rank

EPSS estimates the probability that this vulnerability will be exploited in the wild within the next 30 days. A higher score means more likely to be exploited.

Weakness Type (CWE)

CWE-617 CWE-617

Affected Products

Vendor	Product	Versions
vllm	vllm	0.11.1 — 0.32.0

References

Advisories & Patches

https://github.com/vllm-project/vllm/commit/c15672cfb49152ebf65cfd188e71b84593ec5686 https://github.com/vllm-project/vllm/pull/59417 https://github.com/vllm-project/vllm/security/advisories/GHSA-gq5q-89c7-vhvx

Other References

https://github.com/vllm-project/vllm/releases/tag/v0.32.0

Frequently Asked Questions

What is CVE-2026-115914? +

vLLM is an inference and serving engine for large language models. From 0.11.1 until 0.32.0, the DeepSeek-OCR, DeepSeek-OCR-2, and Unlimited-OCR image path allows a remote client that can reach the OpenAI-compatible API to submit a legal image whose normalized pixels sum to zero when that image is alone in its per-step encoder batch. _parse_and_validate_image_input returns no embedding, embed_multimodal propagates that value, and sanity_check_mm_encoder_outputs triggers a fatal assertion that shuts down the engine for all users. This issue is fixed in version 0.32.0. It has a CVSS v3.1 base score of 6.5 (MEDIUM).

How severe is CVE-2026-115914? +

CVE-2026-115914 has a CVSS v3.1 score of 6.5 out of 10, rated MEDIUM. This is a medium-severity vulnerability that should be remediated as part of regular maintenance. The EPSS score is 0.0040, placing it in the 0th percentile for exploitation probability.

What products are affected by CVE-2026-115914? +

CVE-2026-115914 affects products from vllm, specifically: vllm. Check the affected products table above for specific version ranges.

How do I check if I'm vulnerable to CVE-2026-115914? +

You can use Secably's free Website Scanner to check your website for known vulnerabilities. For infrastructure scanning, use the Port Scanner to identify exposed services that may be affected. Check the vendor advisories linked above for specific patch and version information.

Related Vulnerabilities

CVE-2026-13058

An authenticated user with basic write privileges can cause the mongod process to terminate abnormally by sending a crafted transaction …

CVE-2025-4321

In a Bluetooth device, using RS9116-WiseConnect SDK experiences a Denial of Service, if it receives malformed L2CAP packets, only hard …

CVE-2025-34458

wb2osz/direwolf (Dire Wolf) versions up to and including 1.8, prior to commit 3658a87, contain a reachable assertion vulnerability in the …

CVE-2026-29116

A vulnerability has been found in some Dahua products could allow an unauthenticated remote attacker to send a specially crafted …

CVE-2026-35058

Improper validation of packet length during tls-crypt-v2 key extraction in OpenVPN 2.6.0 through 2.6.19 and 2.7_alpha1 through 2.7.1 allows authenticated …

CVE-2026-29115

A vulnerability has been found in some Dahua products could allow an authenticated remote attacker to send a specially crafted …

CVE-2026-115914