CVE-2025-48956

vLLM is an inference and serving engine for large language models (LLMs). From 0.1.0 to before 0.10.1.1, a Denial of Service (DoS) vulnerability can be triggered by sending a single HTTP GET request with an extremely large header to an HTTP endpoint. This results in server memory exhaustion, potentially leading to a crash or unresponsiveness. The attack does not require authentication, making it exploitable by any remote user. This vulnerability is fixed in 0.10.1.1.

CVSS v3 7.5 HIGH

7.5^/10

CVSS v3 : HIGH

Vector :

Exploitability : 3.9 / Impact : 3.6

Attack Vector NETWORK

Attack Complexity LOW

Privileges Required NONE

User Interaction NONE

Confidentiality Impact NONE

Integrity Impact NONE

Availability Impact HIGH

Scope UNCHANGED

References

Link	Resource
https://github.com/vllm-project/vllm/commit/d8b736f913a59117803d6701521d2e4861701944	Patch
https://github.com/vllm-project/vllm/pull/23267	Issue Tracking Patch
https://github.com/vllm-project/vllm/security/advisories/GHSA-rxc4-3w6r-4v47	Vendor Advisory

Configurations

Configuration 1 (hide)

cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*

History

No history.

Information

Published : 2025-08-21 15:15

Updated : 2025-10-09 18:04

NVD link : CVE-2025-48956

Mitre link : CVE-2025-48956

CVE.ORG link : CVE-2025-48956

JSON object : View

Products Affected

vllm

vllm

CWE

CWE-400

Uncontrolled Resource Consumption

{"id": "CVE-2025-48956", "cveTags": [], "metrics": {"cvssMetricV31": [{"type": "Secondary", "source": "[email protected]", "cvssData": {"scope": "UNCHANGED", "version": "3.1", "baseScore": 7.5, "attackVector": "NETWORK", "baseSeverity": "HIGH", "vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H", "integrityImpact": "NONE", "userInteraction": "NONE", "attackComplexity": "LOW", "availabilityImpact": "HIGH", "privilegesRequired": "NONE", "confidentialityImpact": "NONE"}, "impactScore": 3.6, "exploitabilityScore": 3.9}]}, "published": "2025-08-21T15:15:32.230", "references": [{"url": "https://github.com/vllm-project/vllm/commit/d8b736f913a59117803d6701521d2e4861701944", "tags": ["Patch"], "source": "[email protected]"}, {"url": "https://github.com/vllm-project/vllm/pull/23267", "tags": ["Issue Tracking", "Patch"], "source": "[email protected]"}, {"url": "https://github.com/vllm-project/vllm/security/advisories/GHSA-rxc4-3w6r-4v47", "tags": ["Vendor Advisory"], "source": "[email protected]"}], "vulnStatus": "Analyzed", "weaknesses": [{"type": "Primary", "source": "[email protected]", "description": [{"lang": "en", "value": "CWE-400"}]}], "descriptions": [{"lang": "en", "value": "vLLM is an inference and serving engine for large language models (LLMs). From 0.1.0 to before 0.10.1.1, a Denial of Service (DoS) vulnerability can be triggered by sending a single HTTP GET request with an extremely large header to an HTTP endpoint. This results in server memory exhaustion, potentially leading to a crash or unresponsiveness. The attack does not require authentication, making it exploitable by any remote user. This vulnerability is fixed in 0.10.1.1."}, {"lang": "es", "value": "vLLM es un motor de inferencia y servicio para modelos de lenguaje grandes (LLM). Desde la versi\u00f3n 0.1.0 hasta versiones anteriores a la 0.10.1.1, se pod\u00eda activar una vulnerabilidad de denegaci\u00f3n de servicio (DoS) al enviar una sola solicitud HTTP GET con un encabezado extremadamente grande a un endpoint HTTP. Esto provocaba el agotamiento de la memoria del servidor, lo que podr\u00eda provocar un bloqueo o la falta de respuesta. El ataque no requiere autenticaci\u00f3n, por lo que cualquier usuario remoto puede explotarlo. Esta vulnerabilidad se corrigi\u00f3 en la versi\u00f3n 0.10.1.1."}], "lastModified": "2025-10-09T18:04:53.383", "configurations": [{"nodes": [{"negate": false, "cpeMatch": [{"criteria": "cpe:2.3:a:vllm:vllm:*:*:*:*:*:*:*:*", "vulnerable": true, "matchCriteriaId": "1C8A1151-E1D6-4DA9-B229-DB4AF9D7F027", "versionEndExcluding": "0.10.1.1", "versionStartIncluding": "0.1.0"}], "operator": "OR"}]}], "sourceIdentifier": "[email protected]"}