CVE-2026-93989
vLLM through 0.29.0 fails to properly validate bad_words token …
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.
What to do now
- Check your exposureFind out whether the affected software runs anywhere in your environment, including systems of suppliers.
- UpdateInstall the version in which the vendor fixed the vulnerability. The references lead to the advisory.
- Set the orderSystems that can be reached from the internet come first, then the rest.
References
- github.com/vllm-project/vllm
- github.com/vllm-project/vllm/blob/98dff2a81d747d1dba01a47f939f48c3526d4206/vllm/sampling_p
- github.com/vllm-project/vllm/blob/98dff2a81d747d1dba01a47f939f48c3526d4206/vllm/v1/worker/
- github.com/vllm-project/vllm/pull/48824
- www.vulncheck.com/advisories/vllm-through-0.29.0-cross-request-logits-corruption-via-bad-w
- nvd.nist.gov/vuln/detail/CVE-2026-93989
- cve.org/CVERecord?id=CVE-2026-93989
Data: CVE list and the NVD. This product uses data from the NVD API but is not endorsed or certified by the NVD.