{"id":"CVE-2026-116178","title":"vLLM 0.20.0 before 0.30.0 DoS via /inference/v1/generate mm_placeholders","summary":"vLLM from 0.20.0 before 0.30.0 contains a denial of service vulnerability that allows remote attackers to crash the engine by sending inconsistent multimodal metadata to the disaggregated /inference/v1/generate endpoint. Attackers can re…","severity":"medium","cvss":5.9,"cvssVector":"CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H","cvssSource":"cna","cwe":["CWE-20","CWE-1288"],"vendor":"vllm-project","product":"vllm","affected":["vllm >= 0.20.0 < 0.30.0"],"ssvc":{"exploitation":"none","automatable":"no","technicalImpact":"partial","timestamp":"2026-10-26T19:38:11.562668Z"},"published":"2026-10-25","updated":"2026-10-26","sourceUpdated":"2026-10-26T19:48:35.925Z","source":"CVEORG","sourceUrl":"https://www.cve.org/CVERecord?id=CVE-2026-116178","references":[{"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-34qw-q2h2-jp3m","label":"GitHub Security Advisory (GHSA-34qw-q2h2-jp3m)"},{"url":"https://www.vulncheck.com/advisories/vllm-0.20.0-before-0.30.0-dos-via-inference-v1-generate-mm-placeholders","label":"VulnCheck Advisory: vLLM 0.20.0 before 0.30.0 DoS via /inference/v1/generate mm_placeholders"},{"url":"https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-116178.json"},{"url":"https://access.redhat.com/security/cve/CVE-2026-116178"},{"url":"https://bugzilla.redhat.com/show_bug.cgi?id=2560446"},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-116178"},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-116178"}],"tags":["cve.org","csaf","vex","red-hat"],"epss":0.00318,"epssPercentile":0.24591,"ingestedAt":"2026-10-25T22:08:29.946Z","slug":"CVE-2026-116178","body":"## Overview\n\nvLLM from 0.20.0 before 0.30.0 contains a denial of service vulnerability that allows remote attackers to crash the engine by sending inconsistent multimodal metadata to the disaggregated /inference/v1/generate endpoint. Attackers can replay a valid multimodal item with repeated or overlapping mm_placeholders ranges or mismatched kwargs_data and mm_hashes lists, producing a negative text span in Qwen2.5-VL position construction that terminates the engine.\n\n## Affected\n\n- `vllm >= 0.20.0 < 0.30.0`\n\n## Remediation\n\nRefer to the linked advisories for vendor-supplied fixes and affected version ranges.\n\n## Vendor advisories\n\n- **Red Hat VEX** · Moderate · affected: Red Hat AI Inference Server, Red Hat Enterprise Linux AI (RHEL AI) 3, Red Hat OpenShift AI (RHOAI) · no fix planned: Red Hat AI Inference Server, Red Hat Enterprise Linux AI (RHEL AI) 3, Red Hat OpenShift AI (RHOAI) · updated 2026-10-25 · [vex](https://security.access.redhat.com/data/csaf/v2/vex/2026/cve-2026-116178.json)","depth":"sunlit","depthScore":33,"depthScoreParts":{"impact":32.5,"likelihood":0.1,"exploitation":0,"ransomware":0},"changes":[]}