CVE-2026-48746

vLLM: OpenAI auth bypass

State
Published
Severity
9.1 Critical
Published
Jun 22, 2026
Updated
Sep 16, 2026
Reserved
May 22, 2026

Read the record on cve.org

Description

vLLM is an inference and serving engine for large language models (LLMs). From 0.3.0 until 0.22.0, a vulnerability in ASGI web servers and starlette's trust on those web servers enables an authentication bypass of the OpenAI API AuthenticationMiddleware. It allows to use the API without providing the configured VLLM_API_KEY or --api-key. This vulnerability is fixed in 0.22.0.

Scores

CVSSScoreVectorSource
3.19.1 CriticalCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:HGitHub_M
3.19.1 CriticalCVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:Hredhat-SADP

Weaknesses

  • CWE-444 Inconsistent Interpretation of HTTP Requests ('HTTP Request/Response Smuggling')
  • CWE-501 Trust Boundary Violation

Affected Products

ProductVersions
vllmvllm-project
  • >= 0.3.0, < 0.22.0Affected

References

Every record here comes from the CVE Program’s CVE List. © 1999–2026 The MITRE Corporation. CVE is a registered trademark of The MITRE Corporation. The records are reproduced under the CVE Program terms of use.