CVE-2025-62372
mediumPriorität · verfolgen · 26EPSS 0.004 · 35. PerzentilDEEURUUSEPSS · bewertet 2026-10-27
vLLM is an inference and serving engine for large language models (LLMs). From version 0.5.5 to before 0.11.1, 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). This issue has been patched in version 0.11.1.
cvss
Technischer Schweregrad, 0–10. Eine Zeile je Quelle.
Basiswert 0–10 · höher = schwerwiegender · Vektor überfahren zum Auflösen⚠ Quellen uneinig · medium–high · 5.7–8.3
SSVC
Eine Handlungsentscheidung — aus Ausnutzung und Auswirkung.
schwächen
Die CWE-Schwächeklassen, zu denen dies gehört.
betroffene Produkte (10)
Betroffene Produkte und Versionsbereiche.
Advisories (3)
Hersteller- und Ökosystem-Meldungen dazu.
- WID-SEC-W-2026-0190vllm: Mehrere SchwachstellenDeutsch2026-01-21csaf · cert_bundDE
Ein entfernter, authentisierter Angreifer kann mehrere Schwachstellen in vllm ausnutzen, um einen Denial of Service Angriff durchzuführen oder Code auszuführen.
- GHSA-pmqf-x6x8-p7qwvLLM vulnerable to DoS with incorrect shape of multimodal embedding inputs2025-11-20ghsa · osvUS
### 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: - For models that support image embedding inputs, the engine crashes when scattering the embeddings to `inputs_embeds` (mismatched shape) - For models that don't support image embedding inputs, the engine crashes when validating the inputs inside `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 - Denial of service by crashing the engine ### Mitigation - Use API key to limit access to trusted users. - Set `--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 - https://github.com/vllm-project/vllm/pull/27204
- PYSEC-2026-2019vLLM vulnerable to DoS with incorrect shape of multimodal embedding inputs2026-07-07vendor · osvUS
### 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: - For models that support image embedding inputs, the engine crashes when scattering the embeddings to `inputs_embeds` (mismatched shape) - For models that don't support image embedding inputs, the engine crashes when validating the inputs inside `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 - Denial of service by crashing the engine ### Mitigation - Use API key to limit access to trusted users. - Set `--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 - https://github.com/vllm-project/vllm/pull/27204
auch bekannt als (7)
Weitere Namen und Kennungen dieser Entität.
quellen
Welche Quellen diesen Datensatz belegen und wann.
Referenzen (13)
Externe Belege — Patches, Meldungen, Analysen.
- https://github.com/vllm-project/vllm/commit/58fab50d82838d5014f4a14d991fdb9352c9c84bcnnvd
- https://github.com/vllm-project/vllm/pull/27204cnnvd
- https://github.com/vllm-project/vllm/pull/6613cnnvd
- https://github.com/vllm-project/vllm/releasesPatchcnnvd
- https://github.com/vllm-project/vllm/security/advisories/GHSA-pmqf-x6x8-p7qwcnnvd
- https://github.com/vllm-project/vllm/commit/58fab50d82838d5014f4a14d991fdb9352c9c84benisa_euvd
- https://github.com/vllm-project/vllm/pull/27204enisa_euvd
- https://github.com/vllm-project/vllm/pull/6613enisa_euvd
Angreifer nutzen Schwachstellen wie diese gegen im Internet exponierte Dienste aus. Sehen Sie mit einem kostenlosen Passiv-Scan, was exponiert ist.
Kostenloser Scan