{"dataType":"CVE_RECORD","dataVersion":"5.2","cveMetadata":{"cveId":"CVE-2026-115916","assignerOrgId":"a0819718-46f1-4df5-94e2-005712e83aaa","state":"PUBLISHED","assignerShortName":"GitHub_M","dateReserved":"2026-10-23T20:21:45.107Z","datePublished":"2026-10-26T17:18:17.582Z","dateUpdated":"2026-10-26T17:18:17.582Z"},"containers":{"cna":{"title":"vLLM: Out-of-vocabulary prompt token id passes input validation and triggers a CUDA device-side assert (engine crash / DoS)","problemTypes":[{"descriptions":[{"cweId":"CWE-129","lang":"en","description":"CWE-129: Improper Validation of Array Index","type":"CWE"}]},{"descriptions":[{"cweId":"CWE-193","lang":"en","description":"CWE-193: Off-by-one Error","type":"CWE"}]}],"metrics":[{"cvssV3_1":{"attackComplexity":"LOW","attackVector":"NETWORK","availabilityImpact":"HIGH","baseScore":6.5,"baseSeverity":"MEDIUM","confidentialityImpact":"NONE","integrityImpact":"NONE","privilegesRequired":"LOW","scope":"UNCHANGED","userInteraction":"NONE","vectorString":"CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H","version":"3.1"}}],"references":[{"name":"https://github.com/vllm-project/vllm/security/advisories/GHSA-q97c-p368-75c4","tags":["x_refsource_CONFIRM"],"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-q97c-p368-75c4"},{"name":"https://github.com/vllm-project/vllm/pull/59491","tags":["x_refsource_MISC"],"url":"https://github.com/vllm-project/vllm/pull/59491"},{"name":"https://github.com/vllm-project/vllm/commit/37cecc8141e2dcba9dd534897a164d4baaf27229","tags":["x_refsource_MISC"],"url":"https://github.com/vllm-project/vllm/commit/37cecc8141e2dcba9dd534897a164d4baaf27229"},{"name":"https://github.com/vllm-project/vllm/releases/tag/v0.32.0","tags":["x_refsource_MISC"],"url":"https://github.com/vllm-project/vllm/releases/tag/v0.32.0"}],"affected":[{"vendor":"vllm-project","product":"vllm","versions":[{"version":"< 0.32.0","status":"affected"}]}],"providerMetadata":{"orgId":"a0819718-46f1-4df5-94e2-005712e83aaa","shortName":"GitHub_M","dateUpdated":"2026-10-26T17:18:17.582Z"},"descriptions":[{"lang":"en","value":"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."}],"source":{"advisory":"GHSA-q97c-p368-75c4","discovery":"UNKNOWN"}}}}