CVE-2026-115916: vLLM is an inference and serving engine for large language models

Q: How do I fix CVE-2026-115916?

vllm: upgrade to 0.32.0. Exact commands for each OS are on this page.

Q: Is CVE-2026-115916 being exploited?

CVE-2026-115916 is not on the CISA Known Exploited Vulnerabilities list at the time of the last update. Check again, as this can change.

Q: What does CVE-2026-115916 affect?

CVE-2026-115916 affects vllm. Severity is medium (CVSS 6.5). vLLM is an inference and serving engine for large language models. Prior to 0.32.0, vllm/tokenizers/hf.py computes max_token_id with an off-by-one upper bound, allowing a pre-tokenized prompt ID equal to the vocabulary size to pass validati

Q: How urgent is CVE-2026-115916?

Priority 29/100. No elevated risk signals yet.

Routine

MEDIUM 6.5UPGRADE

CVE-2026-115916 is a medium-severity flaw (CVSS 6.5) affecting vllm. Below: why it ranks where it does, which versions are affected, what to do, and the exact commands.

vLLM is an inference and serving engine for large language models. Prior to 0.32.0, vllm/tokenizers/hf.py computes max_token_id with an off-by-one upper bound, allowing a pre-tokenized prompt ID equal to the vocabulary size to pass validation. For a tensor-parallel size of one and an exact multiple-of-64 vocabulary, VocabParallelEmbedding.forward does not mask that value before F.embedding uses it, causing a CUDA device-side assertion and termina

Why this priority

No elevated risk signals yet

A fixed release or distro patch exists.

Risk facts

CVSS score	6.5 (Medium)
Attack vector	Network (remote)
Attack complexity	Low
Privileges needed	Low
User interaction	None
EPSS exploit chance	0.3% (top 77%)
CISA exploited list	No
Weakness	CWE-129, CWE-193
Published	2026-10-26
Last modified	2026-10-28

Publisher (CNA): GitHub, Inc. · Vendor: vllm

Last updated 2026-10-28 · Sources: CVE Program, NVD, CISA KEV, EPSS and vendor advisories · How we score

Am I affected?

Timeline

2026-10-26 18:17 UTCPublished CVE record published
2026-10-28 15:36 UTCUpdated Record last modified

vllm vllm UPGRADE

Affected: < 0.32.0

Fixed in: 0.32.0

Safe to downgrade to: no unaffected older release

1. Check installed version

apt-cache policy vllm || rpm -q vllm || vllm --version

2. UPGRADE to 0.32.0 (Debian / Ubuntu)

sudo apt update && sudo apt install --only-upgrade vllm

2. UPGRADE (RHEL / Amazon Linux / Fedora)

sudo dnf upgrade vllm

Note

Distros often backport fixes without changing the version number. Confirm with the distro advisory or changelog.

Advisories and coverage

Research and news

Vulners Bulletins: CVE-2026-115916 vLLM: Out-of-vocabulary prompt token id passes input validation and triggers a CUDA device-side assert (engine crash / DoS) 2026-10-26

Verify at the official source NVD · CVE Program · Report a problem with this page
Scores and fixes here are computed automatically from those and other public sources. Confirm with the vendor advisory before changing production systems.

Quick answers

How do I fix CVE-2026-115916?

vllm: upgrade to 0.32.0. Exact commands for each OS are on this page.

Is CVE-2026-115916 being exploited?

CVE-2026-115916 is not on the CISA Known Exploited Vulnerabilities list at the time of the last update. Check again, as this can change.

What does CVE-2026-115916 affect?

CVE-2026-115916 affects vllm. Severity is medium (CVSS 6.5). vLLM is an inference and serving engine for large language models. Prior to 0.32.0, vllm/tokenizers/hf.py computes max_token_id with an off-by-one upper bound, allowing a pre-tokenized prompt ID equal to the vocabulary size to pass validati

How urgent is CVE-2026-115916?

Priority 29/100. No elevated risk signals yet.

Related CVEs

CVE-2026-116172 MEDIUM 5.9 UPGRADE vLLM from 0.22.0 before 0.24.0 contains an improper input validation vulnerability in the Rust gRPC converter
CVE-2026-116173 MEDIUM 5.9 UPGRADE vLLM from 0.22.0 before 0.24.0 contains a denial of service vulnerability that allows remote attackers to term
CVE-2026-116174 MEDIUM 5.9 UPGRADE vLLM from 0.22.0 before 0.24.0 contains an improper input validation vulnerability in the Rust gRPC Generate a
CVE-2026-116175 MEDIUM 6.5 UPGRADE vLLM from 0.19.0 before 0.24.0 contains a decompression bomb vulnerability because the AudioMediaIO multimodal
CVE-2026-116176 HIGH 8.2 UPGRADE vLLM from 0.9.2 before 0.24.0 contains a server-side request forgery vulnerability that allows API callers to

References

← Back to latest CVEs