CVE-2026-54234

[CVE catalog]

Severity

High · CVSS 7.5CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Published

Jul 1, 2026, 10:00:00 PM UTC

Source

[NVD] · [CVE record]

EPSS

0.62% · Oct 25, 2026

A flaw was found in vLLM, a high-throughput and memory-efficient inference and serving engine for Large Language Models (LLMs). A remote attacker can exploit this vulnerability by sending a specially crafted multi-request speculative decoding workload through public gRPC Generate and Abort endpoints. This malformed workload can cause the rejection sampler to produce an out-of-vocabulary token, which then crashes the engine worker. This leads to a service-wide Denial of Service (DoS) for all clients until the worker is restarted. vllm: vLLM: Denial of Service via malformed speculative decoding workload This Important denial of service flaw in vLLM, utilized by Red Hat AI Inference Server, Red Hat Enterprise Linux AI, and Red Hat OpenShift AI, allows a remote attacker to crash the engine worker. By sending a specially crafted multi-request speculative decoding workload to public gRPC endpoints, an attacker can trigger an out-of-vocabulary token, leading to a service-wide disruption for all connected clients. The CVSS score(s) listed for this vulnerability do not reflect the associated product's status, and are included for informational purposes to better understand the severity of this vulnerability.