CVE Explore

Improper input validation in Vllm

CVE-2026-54234

CVSS v3
7.5 High
Public exploits
None known
EPSS
0.006 · 50.5th percentile
Published
2026-07-06

What to do

Fixed in
0.24.0
Affected
2 affected products

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.

Published · last modified .

CVSS v3 metric

CVSS v3 base score 7.5 (High). Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H.

Attack Vector
Network AV:N
Attack Complexity
Low AC:L
Privileges Required
None PR:N
User Interaction
None UI:N
Scope
Unchanged S:U
Confidentiality
None C:N
Integrity
None I:N
Availability
High A:H

EPSS exploit prediction

EPSS 0.006, 50.5th percentile (last changed ); unchanged over 30 days. About EPSS.

Affected products

Weakness classification (CWE)

References