CVE-2026-73558 vLLM: Cross-User Data Leak Vulnerability
An integer overflow in vLLM allows attackers to read other users' inference results via batched requests.
- Exploited in the wild
- Not known
- Login needed
- Not needed
- Attackable
- Over the internet
- Victim action needed
- A user has to click or open something
- Fix
- Available
What is CVE-2026-73558?
CVE-2026-73558 is a medium-severity vulnerability in Vllm Vllm, published on 13 Aug 2026. An attacker can exploit it over the internet without logging in, and only if a user opens a file or follows a link.
Official description
vLLM is an inference and serving engine for large language models. Prior to 0.27.0, an integer overflow in blockIdx.x * 2 * d in activation_kernels.cu can cause act_and_mul_kernel to consume another batched user's input, allowing a request processed in the same inference batch to receive a partial or complete copy of another user's inference result. This issue is fixed in version 0.27.0.
Weakness: CWE-190
Which versions are affected by CVE-2026-73558?
3 open source packages ship an affected version, including vllm. It affects 1 product.
| Package | Affected | Fixed in |
|---|---|---|
vllm PyPI | All versions | 0.27.0 |
vllm PyPI | All versions | 0.27.0 |
vllm PyPI | All versions | 0.27.0 |
Products include Vllm Vllm.
How do I fix CVE-2026-73558?
Upgrade vllm to 0.27.0 or later.
Is CVE-2026-73558 being exploited?
It is not confirmed as exploited in the wild, but public exploit code for CVE-2026-73558 exists.
Other vulnerabilities in Vllm Vllm
- CVE-2025-32444 vLLM Vulnerable to Remote Code Execution via Mooncake Integration
- CVE-2026-22778 vLLM leaks a heap address when PIL throws an error
- CVE-2025-47277 vLLM Allows Remote Code Execution via PyNcclPipe Communication Service
- CVE-2024-11041 Remote Code Execution in vllm-project/vllm
- CVE-2026-48746 vLLM: OpenAI auth bypass
Sources
Also known as GHSA-7m6h-x95x-82q5, MINI-5654-wgg2-f6c6, MINI-7hhv-gr99-276x, MINI-qmmq-rx4c-rjgq, PYSEC-2026-3935.