CVE-2026-48746: vLLM OpenAI API Authentication Bypass Vulnerability
Summary
CVE-2026-48746: vLLM: OpenAI auth bypass. Affects vllm versions >= 0.3.0, < 0.22.0. Fixed in 0.22.0. How to fix ↓
Check if this vulnerability affects your application
Upload a manifest or SBOM and we'll show whether you're affected, plus every other risky package in the same scan.
No signup required · Results in about 60 seconds
Background
vLLM is an open-source project that provides a high-performance implementation of the OpenAI API. It allows developers to interact with OpenAI's models efficiently. The package is designed to handle requests and responses seamlessly, but it has a critical vulnerability that can allow unauthorized access to the API.
Vulnerability Detail
A vulnerability in ASGI web servers and starlette's trust on those web servers enables an authentication bypass of the OpenAI API 'AuthenticationMiddleware'. This issue arises from the way the URL is reconstructed from the request scope, allowing attackers to manipulate the 'Host:' header. By including special URL characters, an attacker can bypass the API key requirement and access the API without proper authentication. This vulnerability affects instances of vLLM that expose the API to potential attackers, especially those not behind a secure web server.
How To Fix
Latest published version: 0.32.0 as of 22 Oct 2026
pip install --upgrade "vllm==0.32.0"Upgrading to it may include breaking changes; the version below is the minimum fixed release for this vulnerability.
If you need the minimum change
0.22.0
This is the first release that contains the fix.
pip install --upgrade "vllm==0.22.0"Tip — Not sure which version your app runs? Vulert can match this fix to it. Scan free →
Workaround
If you cannot immediately update the package, consider implementing a reverse proxy with strict filtering on the 'Host:' header to prevent unauthorized access. This can be done using a web server like Nginx or Apache to validate incoming requests before they reach the vLLM application.
Conclusion
The vLLM package has a critical vulnerability that allows for authentication bypass of the OpenAI API. It is essential to update to the latest version to mitigate this risk. Additionally, implementing a secure web server can help protect against unauthorized access.
See every vulnerability recorded for vllm →
FAQ
The CVSS score of this vulnerability is 9.1, indicating a critical severity level.
You can use Vulert's playground at Vulert Scanner to check if your application is affected by this vulnerability without needing to sign up.
The potential impacts include unauthorized access to the OpenAI API, which could lead to misuse of the API and exposure of sensitive data.
Instances of vLLM that are behind an RFC-conforming web server, such as Nginx, are not affected by this vulnerability.
If you cannot update the package, implement a reverse proxy with strict header filtering as a temporary workaround.
References
Related vulnerabilities
Recent vulnerabilities recorded for vllm.
