vllm
Affected
- introduced
0.5.5, fixed0.11.1(ecosystem)
Fixed versions
0.11.1
Fix
- Fixed version
0.11.1pip install --upgrade vllm==0.11.1
Advisory
Also known as: CVE-2025-62372, GHSA-pmqf-x6x8-p7qw
vLLM vulnerable to DoS with incorrect shape of multimodal embedding inputs
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:
inputs_embeds (mismatched shape)get_input_embeddings (validation fails).This happens because we only validate ndim of the tensor, but not the full shape, in input processor (via MultiModalDataParser).
Impact
Mitigation
--limit-mm-per-prompt to 0 for all non-text modalities to ban multimodal inputs, which includes multimodal embedding inputs. However, the model would then only accept text, defeating the purpose of using a multi-modal model.Resolution
Medium Basis: CVSS v3 score 6.5, vector CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
0.5.5, fixed 0.11.1 (ecosystem)0.11.1
0.11.1pip install --upgrade vllm==0.11.1Fix commands are the export's own generic commands. Ojuju never runs them. Check them against your setup first.
Check your own computer: guard packages check
Source: PyPA Advisory Database · CC BY 4.0