{"id":"CVE-2026-116178","aliases":["EUVD-2026-109439","GHSA-34qw-q2h2-jp3m","GHSA-w9hx-wr9v-56wv","PYSEC-2026-4353"],"status":"published","title":"vLLM 0.20.0 before 0.30.0 DoS via /inference/v1/generate mm_placeholders","description":"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 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.","severity":{"version":"4.0","score":8.2,"vector":"CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N","level":"high","assessor":"cna","from":"cve"},"epss":{"day":"2026-11-11","score":0.00331,"percentile":0.26583,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":true},"cwes":["CWE-20"],"published":"2026-10-25T13:20:17Z","modified":"2026-11-10T02:31:08Z","fixes":[{"ecosystem":"PyPI","package":"vllm","fixed_in":["0.30.0"],"sources":["osv"]}],"affected":[{"source":"osv","ecosystem":"PyPI","package":"vllm","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0.20.0"},{"fixed":"0.30.0"}]}],"versions":["0.20.0","0.20.1","0.20.2","0.21.0","0.22.0","0.22.1","0.23.0","0.24.0","0.25.0","0.25.1","0.26.0","0.27.0","0.27.1","0.28.0","0.29.0"]}],"products":[{"vendor":"vllm","product":"vllm"},{"vendor":"vllm-project","product":"vllm"}],"references":[{"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-34qw-q2h2-jp3m","tags":["vendor-advisory","Patch","Vendor Advisory","advisory","fix"]},{"url":"https://www.vulncheck.com/advisories/vllm-0.20.0-before-0.30.0-dos-via-inference-v1-generate-mm-placeholders","tags":["third-party-advisory","Third Party Advisory","advisory"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-116178","tags":["advisory"]},{"url":"https://github.com/advisories/GHSA-w9hx-wr9v-56wv"},{"url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/116xxx/CVE-2026-116178.json","tags":["advisory"]}],"sources":{"cve":[{"id":"CVE-2026-116178","title":"vLLM 0.20.0 before 0.30.0 DoS via /inference/v1/generate mm_placeholders","description":"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 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.","status":"published","published":"2026-10-25T13:20:17.838Z","modified":"2026-10-26T19:48:35.925Z","cvss":[{"version":"4.0","score":8.2,"vector":"CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N","level":"high","assessor":"cna"},{"version":"3.1","score":5.9,"vector":"CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H","level":"medium","assessor":"cna"}],"cwes":["CWE-20"],"updated_at":"2026-11-02T14:20:03.827373Z"}],"euvd":[{"id":"EUVD-2026-109439","description":"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 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.","status":"published","published":"2026-10-25T13:20:17Z","modified":"2026-10-26T19:48:35Z","cvss":[{"version":"4.0","score":8.2,"vector":"CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N","level":"high"}],"updated_at":"2026-11-02T16:59:28.357127Z"}],"ghsa":[{"id":"GHSA-w9hx-wr9v-56wv","title":"vLLM from 0.20.0 before 0.30.0 contains a denial of service vulnerability that allows remote...","description":"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 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.","status":"published","published":"2026-10-25T15:32:19Z","modified":"2026-10-25T15:32:20Z","cvss":[{"version":"4.0","score":8.2,"vector":"CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X","level":"high"},{"version":"3.1","score":5.9,"vector":"CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H","level":"medium"}],"cwes":["CWE-20"],"updated_at":"2026-11-02T15:14:22.728467Z"}],"nvd":[{"id":"CVE-2026-116178","description":"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 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.","status":"published","published":"2026-10-25T14:16:57.483Z","modified":"2026-10-28T15:40:43.623Z","cvss":[{"version":"4.0","score":8.2,"vector":"CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X","level":"high","assessor":"secondary"},{"version":"3.1","score":5.9,"vector":"CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H","level":"medium","assessor":"secondary"},{"version":"3.1","score":7.5,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H","level":"high","assessor":"primary"}],"cwes":["CWE-20"],"updated_at":"2026-11-02T14:13:03.126833Z"}],"osv":[{"id":"CVE-2026-116178","title":"vLLM 0.20.0 before 0.30.0 DoS via /inference/v1/generate mm_placeholders","description":"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 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.","status":"published","published":"2026-10-25T13:20:17.838Z","modified":"2026-11-10T02:31:08.48221037Z","cvss":[{"version":"4.0","score":8.2,"vector":"CVSS:4.0/AV:N/AC:H/AT:P/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N","level":"high"}],"cwes":["CWE-20"],"updated_at":"2026-11-10T02:45:42.584685Z"},{"id":"PYSEC-2026-4353","description":"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 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.","status":"published","published":"2026-10-25T14:16:57.483Z","modified":"2026-10-29T11:00:02.784018523Z","cvss":[{"version":"3.1","score":7.5,"vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H","level":"high"}],"updated_at":"2026-11-02T15:52:57.671918Z"}]}}