THREATOPS
THREAT OPS › Threat News › [NVD] CVE-2026-93989 (LOW 3.1) — 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-f

[NVD] CVE-2026-93989 (LOW 3.1) — 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-f

mednvdPublished 2026-09-19

CVE-2026-93989 CVSS: 3.1 LOW Published: 2026-09-19T23:17:10.043

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.

Indicators of compromise

Original source: https://nvd.nist.gov/vuln/detail/CVE-2026-93989