Skip to content
Fencer Vulns
medium 4.2

CVE-2025-46722 json

vLLM has a Weakness in MultiModalHasher Image Hashing Implementation

published 2025-05-28 17 mo ago modified 2026-08-07 3 mo ago also EUVD-2025-16188 , GHSA-c65p-x677-fgj6 , PYSEC-2025-43

cvss
4.2
version 3.1, CNA
cisa kev
Not listed
not in CISA's catalog
epss
0.3%
EPSS 0.3% chance of exploitation in 30 days, percentile 27%
public exploits
0
none known

Summary

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.

Patch / fix

  • pypi vllm 0.9.0
  • vllm-project vllm 0.9.0
  • vllm vllm 0.9.0

Vendor advisories

What the vendors themselves say about their products: affected, fixed or not affected, and what to do.

red hat product security vex CVE-2025-46722 updated cvss 4.2

vllm: vLLM has a Weakness in MultiModalHasher Image Hashing Implementation

  • Red Hat AI Inference Server rhaiis/vllm-cuda-rhel9, rhaiis/vllm-rocm-rhel9 Red Hat Enterprise Linux AI (RHEL AI) rhelai1/ui-rhel9, rhelai1/bootc-amd-rhel9, rhelai1/bootc-intel-rhel9, rhelai1/bootc-nvidia-rhel9, rhelai1/bootc-azure-amd-rhel9, rhelai1/instructlab-amd-rhel9, +5 more
    affected: workaround, no fix available “Fix deferred”

Red Hat CSAF VEX © Red Hat, Inc., from https://security.access.redhat.com/data/csaf/v2/vex/; changed: reduced to product statuses and remediations. CC BY 4.0 · csaf

CISA SSVC

Exploitation
none
Automatable
no
Technical impact
partial
Assessed
17 mo ago

Exploit prediction

0.34% chance of exploitation in the next 30 days, higher than 27% of all scored CVEs (FIRST EPSS, 2026-11-08).

CVSS scores

assigned by version score vector
CNA 3.1 4.2 CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:L/I:N/A:L
GitHub 3.1 4.2 CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:L/I:N/A:L
NVD 3.1 7.3 CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L

Affected packages

ecosystem package affected fixed in
pypi vllm
>=0.7.0, <0.9.0
<0.9.0
0.9.0

Affected products

  • vllm-project vllm

Classification

Weakness
  • CWE-1288 Improper Validation of Consistency within Input
  • CWE-1023 Incomplete Comparison with Missing Factors

References

Sources

Fencer Vulns keeps one entry per vulnerability: the CVE id wins, then GHSA, and records that alias each other are merged, except that two CVEs always keep separate entries. The same data is available as JSON.

CPEs and fixed versions use Anchore's correction of NVD's data: Added CPE configurations because not yet analyzed by NVD.