CVEvulnerability
CVE-2026-115916
6.5 MEDIUM
Official page ↗
- Status
- Analyzed
- Published
- 2026-10-26
- Last modified
- 2026-10-28
- Assigner
- [email protected]
- CVSS 3.1 ([email protected])
- 6.5 MEDIUM — CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
- Weaknesses
- CWE-129, CWE-193
Description
vLLM is an inference and serving engine for large language models. Prior to 0.32.0, vllm/tokenizers/hf.py computes max_token_id with an off-by-one upper bound, allowing a pre-tokenized prompt ID equal to the vocabulary size to pass validation. For a tensor-parallel size of one and an exact multiple-of-64 vocabulary, VocabParallelEmbedding.forward does not mask that value before F.embedding uses it, causing a CUDA device-side assertion and terminating the engine. This issue is fixed in version 0.32.0.
Affected configurations
• vllm vllm (< 0.32.0)
Related
References
- https://github.com/vllm-project/vllm/commit/37cecc8141e2dcba9dd534897a164d4baaf27229 Patch
- https://github.com/vllm-project/vllm/pull/59491 Patch
- https://github.com/vllm-project/vllm/releases/tag/v0.32.0 Release Notes
- https://github.com/vllm-project/vllm/security/advisories/GHSA-q97c-p368-75c4 Patch, Vendor Advisory