CVE-2026-100649 vLLM before 0.29.0 Resource Limit Bypass via Sampler Subclass
Unauthenticated attackers can bypass resource limits in vLLM to exhaust GPU memory, causing denial of service.
- Exploited in the wild
- Not known
- Login needed
- Not needed
- Attackable
- Over the internet
- Victim action needed
- Not needed
- Fix
- Available
What is CVE-2026-100649?
CVE-2026-100649 is a medium-severity denial of service vulnerability in Vllm Vllm, published on 26 Sep 2026. An attacker can exploit it over the internet without logging in, with no action from the victim.
Official description
vLLM before 0.29.0 contains a resource-limit bypass vulnerability in PyNvVideoCodec decoder allocation where sampler subclass shadowing allows independent counter increments. Unauthenticated attackers can select different sampler subclasses in video requests to exceed configured decoder limits and exhaust unaccounted GPU memory.
Weakness: CWE-770
Which versions are affected by CVE-2026-100649?
It affects 1 product.
Products include Vllm Vllm.
How do I fix CVE-2026-100649?
Upgrade to a fixed version.
Upgrade, then put limits on request size, rate and nesting depth so one client cannot exhaust the service. More about Denial of service.
Is CVE-2026-100649 being exploited?
There is no known exploitation of CVE-2026-100649 in the wild and no public exploit code.
Other vulnerabilities in Vllm Vllm
- CVE-2025-32444 vLLM Vulnerable to Remote Code Execution via Mooncake Integration
- CVE-2026-22778 vLLM leaks a heap address when PIL throws an error
- CVE-2025-47277 vLLM Allows Remote Code Execution via PyNcclPipe Communication Service
- CVE-2024-11041 Remote Code Execution in vllm-project/vllm
- CVE-2026-48746 vLLM: OpenAI auth bypass
Sources
Also known as GHSA-hf74-6fwg-ghj7, GHSA-j682-9xp5-rrf3, MINI-24pr-5qc2-rp4j, MINI-489v-gg54-9rhm, MINI-r3xf-6mv6-g7mr.