CVE-2026-47475 Denial of service in TensorRT-LLM
An attacker can cause a denial of service in NVIDIA TensorRT-LLM by triggering a reachable assertion in the sampler thread via the OpenAI-compatible API.
- Exploited in the wild
- Not known
- Login needed
- Not needed
- Attackable
- Local or same network only
- Victim action needed
- Not needed
- Fix
- None listed
What is CVE-2026-47475?
CVE-2026-47475 is a medium-severity denial of service vulnerability in NVIDIA TensorRT-LLM, published on 14 Jul 2026. It can only be exploited from the same machine or local network, with no action from the victim.
Official description
NVIDIA TensorRT-LLM contains a vulnerability in the OpenAI-compatible inference API where an attacker could trigger a reachable assertion in the sampler thread. A successful exploit of this vulnerability might lead to denial of service.
Weakness: CWE-617
Which versions are affected by CVE-2026-47475?
How do I fix CVE-2026-47475?
No fix has been published for CVE-2026-47475 yet. Limit exposure of the affected software until one is available.
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-2026-47475 being exploited?
There is no known exploitation of CVE-2026-47475 in the wild and no public exploit code.
Sources
Also known as GHSA-7f2c-g8wp-6p7v.