cvedb.io
CVE-2026-54234
HIGH · CVSS 7.5
EPSS exploitation probability: 0%
Published 2026-07-06T21:16:56.477 · Last modified 2026-07-06T21:16:56.477

Summary

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

Does this affect you?

Add your gear to cvedb and we'll alert you only when a vendor you run ships something exploited.

Check my exposure →

References

This product uses data from the NVD API but is not endorsed or certified by the NVD. Informational only; not professional security advice.