Medium severity · published Sep 19, 2026
CVE-2026-93989
vLLM through 0.29.0 Cross-Request Logits Corruption via bad_words
4.3
CVSS
0.2%
EPSS
No
On KEV
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 vector
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N
Attack vector
Network
Complexity
Low
Privileges
Low
User interaction
None
Confidentiality
None
Integrity
Low
Availability
None
Scope
Unchanged
Exploitation
Is it being used?
Exploit prediction (EPSS)0.2% · higher than 10% of CVEs
Exploited in the wildNo evidence yet
CISA KEVNot listed
Ransomware campaign useNot known
Affected products
Which versions
VL
vllm · vllm-project
through 0.29.0 · source: vendor, NVD
References
NVD
https://nvd.nist.gov/vuln/detail/CVE-2026-93989
CVE.org
https://www.cve.org/CVERecord?id=CVE-2026-93989
Product
https://github.com/vllm-project/vllm
Product
https://github.com/vllm-project/vllm/blob/98dff2a81d747d1dba01a47f939f48c3526d4206/vllm/sampling_params.py#L694-L753
Product
https://github.com/vllm-project/vllm/blob/98dff2a81d747d1dba01a47f939f48c3526d4206/vllm/v1/worker/gpu/sample/bad_words.py
Issue Tracking
https://github.com/vllm-project/vllm/pull/48824
Third Party Advisory
https://www.vulncheck.com/advisories/vllm-through-0.29.0-cross-request-logits-corruption-via-bad-words
More from vllm
9.8
CVE-2026-25960
SSRF Protection Bypass in vLLM
9.8
CVE-2024-9053
vllm-project vllm version 0.6.0 contains a vulnerability in the AsyncEngineRPCServer() RPC server entrypoints
9.1
CVE-2026-48746
vLLM: OpenAI auth bypass
8.8
CVE-2026-54232
vLLM: Dependency Confusion Vulnerability in vLLM Dockerfile
8.8
CVE-2026-27893
vLLM's hardcoded trust_remote_code=True in NemotronVL and KimiK25 bypasses user security opt-out