CVE-2026-78684: vLLM GPU Backend Misclassification Leading to Denial of Service
Summary
CVE-2026-78684: vLLM before 0.27.0 fails to properly classify DeepStream as a GPU backend and omits pixel-limit enforcement in its decode path. Affects vllm versions < 0.27.0. Fixed in 0.27.0. 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 library designed for efficient language model serving, leveraging GPU resources to enhance performance. It allows users to deploy large language models with minimal latency. However, a recent vulnerability has been identified that affects its ability to properly classify DeepStream as a GPU backend, which can lead to resource management issues.
Vulnerability Detail
The vulnerability arises from vLLM versions prior to 0.27.0, which fail to classify DeepStream correctly as a GPU backend. This misclassification results in the omission of pixel-limit enforcement in the decode path. Unauthenticated attackers can exploit this flaw by activating DeepStream at request time, which initializes a process-wide GPU decode pool. Consequently, this can lead to resource exhaustion and partial denial of service for concurrent requests, impacting the availability of the service.
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
0.27.0
This is the first release that contains the fix.
pip install --upgrade "vllm==0.27.0"Tip — Not sure which version your app runs? Vulert can match this fix to it. Scan free →
Workaround
As a temporary workaround, consider implementing strict access controls to limit unauthenticated access to the vLLM service, thereby reducing the risk of exploitation until the package can be updated.
Conclusion
In summary, the vLLM package has a critical vulnerability that affects its GPU backend classification, leading to potential denial of service. Users are strongly advised to update to version 0.27.0 or later to mitigate this risk. Implementing access controls can serve as a temporary workaround until the update is applied.
See every vulnerability recorded for vllm →
FAQ
The vulnerability can lead to partial denial of service, allowing attackers to exhaust GPU resources and affect the performance of concurrent requests.
You can use Vulert's playground at Vulert Scanner to check if your application is affected by this vulnerability without requiring any signup.
All versions of vLLM prior to 0.27.0 are affected by this vulnerability.
Yes, implementing strict access controls can help mitigate the risk of exploitation until you can update to the patched version.
You can find more information about vLLM and its functionalities on its official GitHub repository.
Related vulnerabilities
Recent vulnerabilities recorded for vllm.
