Vulnerability · Published Sep 19, 2026 · Assigned by VulnCheck

CVE-2026-93989vLLM through 0.29.0 Cross-Request Logits Corruption via bad_words

Low 3.1

In plain English

A low-severity security flaw in vllm from vllm-project. There are no confirmed reports of exploitation so far. The chance it is exploited in the next 30 days is estimated at 0.2% (higher than 10% of all known vulnerabilities). Fixed versions are available - upgrade the affected package.

Severity (CVSS 3.1)3.1Low

Exploit odds (30 days)0.2%top 89.8% of all CVEs

Exploited in the wildNo reportsCISA SSVC: none

Attack toolingNoneno public tooling known

Official description

vLLM through 0.29.0 fails to properly validate bad_words token indices against the model's generation output width in SamplingParams.update_from_tokenizer(). Attackers can supply out-of-bounds token indices that corrupt logits memory of concurrent requests, causing different in-flight HTTP requests to return incorrect tokens.

Affected packages & fixed versions

Ecosystem	Package	Fixed in	Advisory
PyPI	vllm	0.30.0	PYSEC-2026-4000

Affected products (as reported)

Vendor	Product	Versions
vllm-project	vllm	0 – through 0.29.0

Weakness: CWE-129 · Vector: CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:N

References

GitHub PR #48824 issue-tracking, patch
_bad_words_kernel at v0.29.0 technical-description
update_from_tokenizer at v0.29.0 technical-description
https://github.com/vllm-project/vllm product
VulnCheck Advisory: vLLM through 0.29.0 Cross-Request Logits Corruption via bad_words third-party-advisory