CVE-2026-93989 — vLLM through 0.29.0 Cross-Request Logits Corruption via bad_words
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 suppl…
Published: 2026-09-19 · CVSS: 2.3 · CVSS 4.0
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.
Source: CVE Program · Revision 8b16575c39f1e1810c5e863010da073d2d3285e7