CVE-2026-54234: vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Prior to 0.24.0, a frontend-legal multi-request speculative decoding workload can cause the re

highCVSS 7.5CVE-2026-54234

vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Prior to 0.24.0, a frontend-legal multi-request speculative decoding workload can cause the rejection sampler to produce a recovered token equal to the model vocabulary size boundary value, which is then converted to negative one when the engine selects the next live token for a request and is written back into the drafter's input ids; that out-of-vocabulary value is later consumed by the model's embedding and attention path and crashes the engine worker with a GPU device-side assertion. The same triggering request sequence is reachable through the public gRPC Generate and Abort endpoints, so a remote client that can send generation requests can crash the shared engine worker, aborting concurrent requests and causing a service-wide denial of service for other clients of the deployment until the worker is restarted. This issue is fixed in version 0.24.0.

Details

Source: NVD Recent CVEs (US · database · site)
Severity: high — CVSS 7.5
Published: 2026-07-06
Last updated: 2026-07-07
Exploitation: Not in CISA KEV at last sync

Original advisory: https://nvd.nist.gov/vuln/detail/CVE-2026-54234

Exploitation outlook

EPSS (FIRST.org) estimates each CVE’s probability of exploitation in the next 30 days — here is the CSIRTS.com read on those numbers.

Low exploitation riskCVE-2026-542340.34% 30-day exploitation probability — currently an unlikely target, but scores change as exploit code circulates. Riskier than 26% of all scored CVEs.

Referenced CVEs

CVE	CSIRTS overview	External
CVE-2026-54234	coverage & exploitation status	NVD · CVE.org

Same CVEs, other sources

How other CERTs, PSIRTs and databases cover the vulnerabilities in this advisory.

medium[NEW] [medium] vllm: Multiple vulnerabilities allow Denial of Servicecert-bund

Recent advisories for vLLM is a

A cluster of recent advisories against the same product widens the attack surface — attackers routinely chain freshly published CVEs on one product, so review these together.

high[UPDATE] [high] vllm: Vulnerability allows code executioncert-bund · 2026-07-10
medium[NEW] [medium] vllm: Multiple vulnerabilities allow Denial of Servicecert-bund · 2026-07-07
high[UPDATE] [high] vllm: Vulnerability allows bypassing security measurescert-bund · 2026-07-07
medium[UPDATE] [medium] vllm: Vulnerability allows code executioncert-bund · 2026-07-07
medium[UPDATE] [medium] vllm: Multiple vulnerabilitiescert-bund · 2026-07-07
medium[UPDATE] [medium] vllm: Multiple vulnerabilitiescert-bund · 2026-07-07