INCIDENT MAP

Browse / Unbounded Consumption / CVE-2025-62372

vLLM vulnerable to DoS with incorrect shape of multimodal embedding inputs

CVE-2025-62372 ยท 2025-11-20 high

Class
Unbounded Consumption (filed by a model that read the advisory)
Technique
Denial of service
Affected
vllm (PyPI)
CVSS
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
EPSS
0.4% chance of exploitation in 30 days

How it works, in short

Multimodal embedding inputs with the right number of dimensions but a wrong hidden size pass validation and crash the engine during embedding scatter. (Our summary, written by a model. The source text is below.)

Details

Summary

Users can crash the vLLM engine serving multimodal models by passing multimodal embedding inputs with correct ndim but incorrect shape (e.g. hidden dimension is wrong), regardless of whether the model is intended to support such inputs (as defined in the Supported Models page).

The issue has existed ever since we added support for image embedding inputs, i.e. #6613 (released in v0.5.5)

Details

Using image embeddings as an example:

This happens because we only validate ndim of the tensor, but not the full shape, in input processor (via MultiModalDataParser).

Impact

Mitigation

Resolution

References

Source: GitHub Advisory Database, used under CC BY 4.0.