Low severity3.7NVD Advisory· Published Sep 18, 2026· Updated Sep 28, 2026
CVE-2026-93840
CVE-2026-93840
Description
vLLM before 0.29.0 validates allowed_token_ids against tokenizer length instead of model output logits width in SamplingParams._validate_allowed_token_ids(). Attackers can supply token IDs above the output vocabulary that pass validation, causing LogitBiasState to corrupt GPU logits state and allow concurrent requests to sample tokens outside their allowlists.
AI Insight
LLM-synthesized narrative grounded in this CVE's description and references.
Affected products
2Patches
Vulnerability mechanics
References
5- github.com/vllm-project/vllm/commit/5b0e5b69ac1a3884a6479c9537789c95263cc804nvdPatch
- github.com/vllm-project/vllm/pull/49080nvdIssue TrackingPatch
- www.vulncheck.com/advisories/vllm-before-0.29.0-cross-request-logits-corruption-via-allowed-token-idsnvdThird Party Advisory
- github.com/vllm-project/vllm/blob/v0.28.0/vllm/sampling_params.pynvdProduct
- github.com/vllm-project/vllm/blob/v0.28.0/vllm/v1/worker/gpu/sample/logit_bias.pynvdProduct
News mentions
1- VLLM: Batch of 17 Vulnerabilities Disclosed, Affecting LLM Inference Engine StabilityVypr Intelligence · Sep 21, 2026