4.3

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

Analyzed
AliasesPYSEC-2026-4000
Published 19-09-2026Last modified 22-09-2026

Vulnerability summary

CVE-2026-93989 is a medium (CVSS 4.3) Improper Validation of Array Index vulnerability in vllm-project vllm. EPSS estimates a 0.20% chance of exploitation in the next 30 days, higher than 9.6% of scored vulnerabilities.

Overall risk
17/100
0100
Low · default modelHow is this calculated?
CVSS score
4.3/10
010
Medium · v3.1 (NVD)
EPSS score
0.20%
0%100%
Low · 0.00 pp in 30 days
KEVNot listedNot listed by any tracked catalog.
RansomwareNot flaggedNo ransomware use recorded.
Public exploits0No public exploits recorded

Advisory lineage

Upstream 0 · Downstream 1 · Related 1

Timeline

  1. 19-09-2026
    PublishedVulnerability first disclosed.
  2. 22-09-2026
    Last modifiedVulnerability information updated.

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.

CVSS metrics

5 scores · 3 sources · 2 versions

EPSS trend

Probability

0.20%

Chance of exploitation in the next 30 days

Percentile

10%

Versus all other CVEs

30-day change

No change 0.00 pp

From 0.20% on 22-09-2026

22-09-202622-10-2026

Known exploited

Compare KEV catalogs
CISA KEV

Not in the CISA KEV catalog.

ENISA CNW

Not in the ENISA CNW catalog.

CIRCL

Not in the CIRCL catalog.

Public exploits

0 found

No public exploits indexed.

Weaknesses

CWE · root cause

Affected systems

PyPI1 product
  • vllm< 0.30.0
Vllm1 product
Vllm-project1 product

3 vendors · 3 products

References

5 links