CVE-2025-46722: vllm — Fix

Description

vLLM is an inference and serving engine for large language models (LLMs). In versions starting from 0.7.0 to before 0.9.0, in the file vllm/multimodal/hasher.py, the MultiModalHasher class has a security and data integrity issue in its image hashing method. Currently, it serializes PIL.Image.Image objects using only obj.tobytes(), which returns only the raw pixel data, without including metadata such as the image’s shape (width, height, mode). As a result, two images of different sizes (e.g., 30x100 and 100x30) with the same pixel byte sequence could generate the same hash value. This may lead to hash collisions, incorrect cache hits, and even data leakage or security risks. This issue has been patched in version 0.9.0.

Is your site exposed to CVE-2025-46722?

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

CVSS v3.1 Score

4.2

MEDIUM

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

Weakness Type (CWE)

CWE-1023 CWE-1023

CWE-1288 CWE-1288

Affected Products

Vendor	Product	Versions
vllm	vllm	0.7.0 — 0.9.0

References

Advisories & Patches

https://github.com/vllm-project/vllm/commit/99404f53c72965b41558aceb1bc2380875f5d848 https://github.com/vllm-project/vllm/pull/17378 https://github.com/vllm-project/vllm/security/advisories/GHSA-c65p-x677-fgj6

Frequently Asked Questions

What is CVE-2025-46722? +

vLLM is an inference and serving engine for large language models (LLMs). In versions starting from 0.7.0 to before 0.9.0, in the file vllm/multimodal/hasher.py, the MultiModalHasher class has a security and data integrity issue in its image hashing method. Currently, it serializes PIL.Image.Image objects using only obj.tobytes(), which returns only the raw pixel data, without including metadata such as the image’s shape (width, height, mode). As a result, two images of different sizes (e.g., 30x100 and 100x30) with the same pixel byte sequence could generate the same hash value. This may lead to hash collisions, incorrect cache hits, and even data leakage or security risks. This issue has been patched in version 0.9.0. It has a CVSS v3.1 base score of 4.2 (MEDIUM).

How severe is CVE-2025-46722? +

CVE-2025-46722 has a CVSS v3.1 score of 4.2 out of 10, rated MEDIUM. This is a medium-severity vulnerability that should be remediated as part of regular maintenance.

What products are affected by CVE-2025-46722? +

CVE-2025-46722 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-2025-46722? +

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-2026-92611

In Eclipse Ankaios versions 0.6.0 to before 1.0.4, `LogRule::matches` in the agent control-interface authorizer stops at the first wildcard pattern …

CVE-2026-81376 9.6

Incomplete comparison with missing factors in Visual Studio Code allows an unauthorized attacker to bypass a security feature over a …

CVE-2026-85491 8.8

Catalyst::Seal versions before 0.03 for Perl allow one request to disable a path or route a later one past an …

CVE-2026-24255 7.5

NVIDIA Dynamo for Linux contains a vulnerability in the multimodal embedding cache, where an attacker could cause a hash collision …

CVE-2025-62000 7.1

BullWall Ransomware Containment may not always detect an encrypted file. This issue affects a specific file inspection method that evaluates …

CVE-2026-53839 6.5

OpenClaw before 2026.5.7 contains a hostname validation vulnerability in retry endpoint checks that allows matching hostname prefixes instead of exact …

CVE-2025-46722