Goal Reached Thanks to every supporter — we hit 100%!

Goal: 1000 CNY · Raised: 1359 CNY

100%
Buy Us a Coffee

CVE-2026-69147— vLLM: Request-selected PyNvVideoCodec GPU decode bypasses static VRAM reservation

Quick assessment

Affected
vllm-project vllm
Exploitation
No confirmed in-the-wild exploitation; assess based on exposure
Recommended action
Check the vendor advisory and references for a fixed version. If immediate upgrade is impossible, restrict exposure and increase monitoring.

vLLM 是一个用于大型语言模型（LLM）的推理和服务引擎。在 0.28.0 版本之前，Chat Completions 和 Responses 的请求体可以将 设置为 ，即使启动配置选择了软件解码器， 也会将该选择转发给 。引擎的 逻辑仅根据静态配置来预留解码器内存，因此由请求选择的 后端可能会创建 CUDA 上下文、解码器表面（surfaces）以及解码帧的内存分配，而这些并未从引擎的 KV 缓存预算中扣除。能够向安装了 PyNvVideoCodec 的视频支持型 GPU 部署提交视频请求的攻击者，可以耗尽共享

CVSS 6.5 · Medium EPSS 0.58% · P48

Possible ATT&CK Techniques 1 AI

T1499 · Endpoint Denial of Service

Affected Version Matrix 1

VendorProduct Version RangeStatus
vllm-project vllm < 0.28.0 affected
Get alerts for future matching vulnerabilities Log in to subscribe

I. Basic Information for CVE-2026-69147

Vulnerability Information
Have questions about the vulnerability? See if Shenlong's analysis helps!
View Shenlong Deep Dive ↗

Although we use advanced large model technology, its output may still contain inaccurate or outdated information.Shenlong tries to ensure data accuracy, but please verify and judge based on the actual situation.

Vulnerability Title
vLLM: Request-selected PyNvVideoCodec GPU decode bypasses static VRAM reservation
Source: CVE Program / CVE List V5
Vulnerability Description
vLLM is an inference and serving engine for large language models. Prior to 0.28.0, request bodies for Chat Completions and Responses can set media_io_kwargs.video.video_backend to pynvvideocodec, and MediaConnector.fetch_video forwards that choice to VideoMediaIO even when startup configuration selected a software decoder. The engine's _reserve_mm_ipc_gpu_memory logic budgets decoder memory only from static configuration, so the request-selected VIDEO_LOADER_REGISTRY backend can create a CUDA context, decoder surfaces, and decoded-frame allocations that were not removed from the engine's KV-cache budget. An attacker able to submit video requests to a video-capable GPU deployment with PyNvVideoCodec installed can exhaust shared GPU memory, causing request failures, worker crashes, or denial of service. The first release containing the fix is version 0.28.0.
Source: CVE Program / CVE List V5
CVSS Information
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H
Source: CVE Program / CVE List V5
Vulnerability Type
未加控制的资源消耗(资源穷尽)
Source: CVE Program / CVE List V5

Affected Products

Vendor Product Affected Versions CPE Subscribe
vllm-project vllm < 0.28.0 -

II. Public POCs for CVE-2026-69147

# POC Description Source Link Shenlong Link
AI-Generated POC Premium

No public POC found.

Login to generate AI POC

III. Intelligence Information for CVE-2026-69147

请登录查看更多情报信息。

Patches & Fixes for CVE-2026-69147 (3)

Other References for CVE-2026-69147 (1)

Same Patch Batch · vllm-project · 2026-09-16 · 4 CVEs total

CVE-2026-57173 6.5 MEDIUM vLLM: Unauthenticated audio decompression-bomb DoS in /v1/chat/completions
CVE-2026-92220 5.3 MEDIUM vllm-project vLLM MoRIIO Acknowledgement moriio_connector.py MoRIIOWrapper._handle_release
CVE-2026-92365 4.3 MEDIUM vllm-project vllm thinking_budget_state.py algorithmic complexity

IV. Related Vulnerabilities

Same product: vllm
Same vendor: vllm-project
Same weakness: CWE-400