{"api_version":"1","generated_at":"2026-08-13T17:41:15+00:00","cve":"CVE-2026-73559","urls":{"html":"https://cve.report/CVE-2026-73559","api":"https://cve.report/api/cve/CVE-2026-73559.json","docs":"https://cve.report/api","cve_org":"https://www.cve.org/CVERecord?id=CVE-2026-73559","nvd":"https://nvd.nist.gov/vuln/detail/CVE-2026-73559"},"summary":{"title":"vLLM: Completion prompt lists fan out into unbounded engine requests","description":"vLLM is an inference and serving engine for large language models. From 0.19.0 until 0.26.0, the /v1/completions CompletionRequest.prompt field in vllm/entrypoints/openai/completion/protocol.py accepts an unbounded list[str] or list[list[int]], prompt_to_seq() in vllm/renderers/inputs/preprocess.py and OnlineRenderer.preprocess_completion() in vllm/renderers/online_renderer.py expand every element, and vllm/entrypoints/openai/completion/serving.py creates one engine generator and response slot per prompt, allowing an authenticated API client to exhaust CPU, memory, async scheduling capacity, engine request slots, and response buffering with one request. This issue is fixed in version 0.26.0.","state":"PUBLISHED","assigner":"GitHub_M","published_at":"2026-08-13 16:19:05","updated_at":"2026-08-13 16:19:05"},"problem_types":["CWE-400","CWE-400 CWE-400: Uncontrolled Resource Consumption"],"metrics":[{"version":"3.1","source":"security-advisories@github.com","type":"Secondary","score":"6.5","severity":"MEDIUM","vector":"CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H","data":{"version":"3.1","vectorString":"CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H","baseScore":6.5,"baseSeverity":"MEDIUM","attackVector":"NETWORK","attackComplexity":"LOW","privilegesRequired":"LOW","userInteraction":"NONE","scope":"UNCHANGED","confidentialityImpact":"NONE","integrityImpact":"NONE","availabilityImpact":"HIGH"}},{"version":"3.1","source":"CNA","type":"DECLARED","score":"6.5","severity":"MEDIUM","vector":"CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H","data":{"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":[{"url":"https://github.com/vllm-project/vllm/commit/675f4295cdfe0d870471c2b51bfeca3a68a9569e","name":"https://github.com/vllm-project/vllm/commit/675f4295cdfe0d870471c2b51bfeca3a68a9569e","refsource":"security-advisories@github.com","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://github.com/vllm-project/vllm/releases/tag/v0.26.0","name":"https://github.com/vllm-project/vllm/releases/tag/v0.26.0","refsource":"security-advisories@github.com","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://github.com/vllm-project/vllm/pull/47845","name":"https://github.com/vllm-project/vllm/pull/47845","refsource":"security-advisories@github.com","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-87x5-vmc3-756j","name":"https://github.com/vllm-project/vllm/security/advisories/GHSA-87x5-vmc3-756j","refsource":"security-advisories@github.com","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-73559","name":"CVE Program record","refsource":"CVE.ORG","tags":["canonical"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-73559","name":"NVD vulnerability detail","refsource":"NVD","tags":["canonical","analysis"]}],"affected":[{"source":"CNA","vendor":"vllm-project","product":"vllm","version":"affected >= 0.19.0, < 0.26.0","platforms":[]}],"timeline":[],"solutions":[],"workarounds":[],"exploits":[],"credits":[],"nvd_cpes":[],"vendor_comments":[],"enrichments":{"kev":null,"epss":null,"legacy_qids":[]},"source_records":{"cve_program":{"containers":{"cna":{"affected":[{"product":"vllm","vendor":"vllm-project","versions":[{"status":"affected","version":">= 0.19.0, < 0.26.0"}]}],"descriptions":[{"lang":"en","value":"vLLM is an inference and serving engine for large language models. From 0.19.0 until 0.26.0, the /v1/completions CompletionRequest.prompt field in vllm/entrypoints/openai/completion/protocol.py accepts an unbounded list[str] or list[list[int]], prompt_to_seq() in vllm/renderers/inputs/preprocess.py and OnlineRenderer.preprocess_completion() in vllm/renderers/online_renderer.py expand every element, and vllm/entrypoints/openai/completion/serving.py creates one engine generator and response slot per prompt, allowing an authenticated API client to exhaust CPU, memory, async scheduling capacity, engine request slots, and response buffering with one request. This issue is fixed in version 0.26.0."}],"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"}}],"problemTypes":[{"descriptions":[{"cweId":"CWE-400","description":"CWE-400: Uncontrolled Resource Consumption","lang":"en","type":"CWE"}]}],"providerMetadata":{"dateUpdated":"2026-08-13T15:09:02.493Z","orgId":"a0819718-46f1-4df5-94e2-005712e83aaa","shortName":"GitHub_M"},"references":[{"name":"https://github.com/vllm-project/vllm/security/advisories/GHSA-87x5-vmc3-756j","tags":["x_refsource_CONFIRM"],"url":"https://github.com/vllm-project/vllm/security/advisories/GHSA-87x5-vmc3-756j"},{"name":"https://github.com/vllm-project/vllm/pull/47845","tags":["x_refsource_MISC"],"url":"https://github.com/vllm-project/vllm/pull/47845"},{"name":"https://github.com/vllm-project/vllm/commit/675f4295cdfe0d870471c2b51bfeca3a68a9569e","tags":["x_refsource_MISC"],"url":"https://github.com/vllm-project/vllm/commit/675f4295cdfe0d870471c2b51bfeca3a68a9569e"},{"name":"https://github.com/vllm-project/vllm/releases/tag/v0.26.0","tags":["x_refsource_MISC"],"url":"https://github.com/vllm-project/vllm/releases/tag/v0.26.0"}],"source":{"advisory":"GHSA-87x5-vmc3-756j","discovery":"UNKNOWN"},"title":"vLLM: Completion prompt lists fan out into unbounded engine requests"}},"cveMetadata":{"assignerOrgId":"a0819718-46f1-4df5-94e2-005712e83aaa","assignerShortName":"GitHub_M","cveId":"CVE-2026-73559","datePublished":"2026-08-13T15:09:02.493Z","dateReserved":"2026-08-12T20:53:46.380Z","dateUpdated":"2026-08-13T15:09:02.493Z","state":"PUBLISHED"},"dataType":"CVE_RECORD","dataVersion":"5.2"},"nvd":{"publishedDate":"2026-08-13 16:19:05","lastModifiedDate":"2026-08-13 16:19:05","problem_types":["CWE-400","CWE-400 CWE-400: Uncontrolled Resource Consumption"],"metrics":{"cvssMetricV31":[{"source":"security-advisories@github.com","type":"Secondary","cvssData":{"version":"3.1","vectorString":"CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H","baseScore":6.5,"baseSeverity":"MEDIUM","attackVector":"NETWORK","attackComplexity":"LOW","privilegesRequired":"LOW","userInteraction":"NONE","scope":"UNCHANGED","confidentialityImpact":"NONE","integrityImpact":"NONE","availabilityImpact":"HIGH"},"exploitabilityScore":2.8,"impactScore":3.6}]},"configurations":[]},"legacy_mitre":{"record":{"CveYear":"2026","CveId":"73559","Ordinal":"1","Title":"vLLM: Completion prompt lists fan out into unbounded engine requ","CVE":"CVE-2026-73559","Year":"2026"},"notes":[{"CveYear":"2026","CveId":"73559","Ordinal":"1","NoteData":"vLLM is an inference and serving engine for large language models. From 0.19.0 until 0.26.0, the /v1/completions CompletionRequest.prompt field in vllm/entrypoints/openai/completion/protocol.py accepts an unbounded list[str] or list[list[int]], prompt_to_seq() in vllm/renderers/inputs/preprocess.py and OnlineRenderer.preprocess_completion() in vllm/renderers/online_renderer.py expand every element, and vllm/entrypoints/openai/completion/serving.py creates one engine generator and response slot per prompt, allowing an authenticated API client to exhaust CPU, memory, async scheduling capacity, engine request slots, and response buffering with one request. This issue is fixed in version 0.26.0.","Type":"Description","Title":"vLLM: Completion prompt lists fan out into unbounded engine requ"}]}}}