Vulnerability · Published Oct 5, 2026 · Assigned by GitHub_M
CVE-2026-105753vLLM: Mirrored multimodal IPC caches desync after a rejected request — a later request reusing the same media hash trips a receiver assertion in the engine core
Medium 6.5
In plain English
A medium-severity security flaw 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 0.4% (higher than 37% of all known vulnerabilities). Fixed versions are available - upgrade the affected package.
Official description
vLLM is an inference and serving engine for large language models. Prior to 0.28.0, the default mirrored multimodal LRU cache can commit a media hash in the frontend sender cache during multimodal rendering and before engine admission, while the engine receiver cache never receives the payload if that request is rejected. A later request reusing the same media hash causes MultiModalProcessorSenderCache to send no payload and MultiModalReceiverCache to reach an assertion with the message "Expected a cached item," producing a shared-service availability failure. This issue is fixed in version 0.28.0.
Affected packages & fixed versions
| Ecosystem | Package | Fixed in | Advisory |
|---|---|---|---|
| PyPI | vllm | 0.28.0 | GHSA-ph3r-5jfg-f84f |
| PyPI | vllm | 0.28.0 | PYSEC-2026-4197 |
Affected products (as reported)
| Vendor | Product | Versions |
|---|---|---|
| vllm-project | vllm | < 0.28.0 |
Weakness: CWE-617 · Vector: CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
References
- https://github.com/vllm-project/vllm/security/advisories/GHSA-ph3r-5jfg-f84f x_refsource_CONFIRM
- https://github.com/vllm-project/vllm/pull/46747 x_refsource_MISC
- https://github.com/vllm-project/vllm/pull/51897 x_refsource_MISC
- https://github.com/vllm-project/vllm/commit/396204230423b7cc6798300926b8fa30190d26a9 x_refsource_MISC
- https://github.com/vllm-project/vllm/releases/tag/v0.28.0 x_refsource_MISC