Vulnerability · Published Sep 21, 2026 · Assigned by VulnCheck

CVE-2026-94627vLLM through 0.29.0 GPU KV Cache Leak via Mooncake Transfer ID Collision

High 7.5

In plain English

A high-severity memory leak in vllm from vllm-project. There are no confirmed reports of exploitation so far. The chance it is exploited in the next 30 days is estimated at 0.6% (higher than 50% of all known vulnerabilities).

Severity (CVSS 3.1)7.5High

Exploit odds (30 days)0.6%top 49.8% of all CVEs

Exploited in the wildNo reportsCISA SSVC: none

Attack toolingNoneno public tooling known

Official description

vLLM Mooncake connector through 0.29.0 fails to properly manage GPU KV cache block ownership when concurrent child requests share a single transfer ID in prefill/decode disaggregated deployments. Attackers can trigger GPU memory exhaustion by submitting completion requests with multiple prompts, causing orphaned KV cache blocks to accumulate until process restart and eventually preventing legitimate requests from executing.

Affected products (as reported)

Vendor	Product	Versions
vllm-project	vllm	0 – through 0.29.0

Weakness: CWE-401 · Vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H

References

Pull Request #49796 issue-tracking, patch
https://github.com/vllm-project/vllm/blob/v0.29.0/vllm/distributed/kv_transfer/kv_connector/v1/mooncake/mooncake_connector.py#L1978-L1989 technical-description
https://github.com/vllm-project/vllm product
VulnCheck Advisory: vLLM through 0.29.0 GPU KV Cache Leak via Mooncake Transfer ID Collision third-party-advisory