← Back to search

CVE-2026-94626

7.5 HIGHpublic exploit available

Published 2026-09-21 · Updated 2026-09-21

AI risk analysis

Summary
The flaw allows attackers to allocate unbounded memory by supplying arbitrary tp_size values, potentially leading to a denial of service via kernel OOM-kill.
Exploitability
Exploitation is relatively straightforward with preconditions including access to prefill/decode disaggregated deployments.
Blast radius
If exploited, this could result in significant impact, such as service disruption for affected systems.
Prioritized remediation
Update to a patched version of vLLM greater than 0.29.0 or apply vendor-provided patches immediately.
memory-exhaustiondospatch-available

Analysis generated locally by qwen2.5:7b-instruct (no data left the box). AI-assisted — verify against primary sources before acting.

NVD description

vLLM through 0.29.0 fails to validate the tp_size parameter in kv_transfer_params on OpenAI-compatible completion endpoints, allowing attackers to allocate unbounded memory. Attackers can supply arbitrary tp_size values in prefill/decode disaggregated deployments to exhaust memory and trigger kernel OOM-kill of the decode worker process.

CVSS vector

CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

Weaknesses

CWE-789

Public exploit & PoC references

All references

Source data: NVD (nvd.nist.gov), public domain. Exploit-DB.ai adds local AI analysis for defensive use only.

Related CVEs

Related by shared AI tags and CWE weakness class. Browse the full archive.