# CVE-2026-24233: NVIDIA TensorRT-LLM deserialization in model weight unpickler

Source: https://junglewise.ai/threats/cve-2026-24233-nvidia-tensorrt-llm-deserialization-in-model-weight-unpickler  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-07-14. Updated: 2026-07-14  
License: CC BY 4.0 with a link back

CVE-2026-24233 · Severity: high · CVSS 8.4 · Published 2026-07-14

NVIDIA TensorRT-LLM, a library used to optimize and run large language models on Linux, contains a security flaw in how it handles model files. An attacker with local access to the system could provide a malicious model file that, when loaded, allows them to take control of the system. This could result in the theft of sensitive data, unauthorized changes to software, or a complete system takeover.

Technologies: [Nvidia TensorRT-LLM](https://junglewise.ai/threats/technologies/tensorrt-llm). Vendors: [Nvidia](https://junglewise.ai/threats/vendors/nvidia).

## Quick answers

**Is CVE-2026-24233 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-24233 being exploited in the wild.

**How severe is CVE-2026-24233?** High severity, with a CVSS score of 8.4, published 14 July 2026.

**Which products and versions are affected?** Nvidia TensorRT-LLM v1.3.0 rc14 and earlier.

**Is there a fix for CVE-2026-24233?** No fixed version is recorded yet. Check the vendor's advisory in the references below. If you run it, Junglewise drafts a mitigation that fits your setup and applies it once you approve, with senior security engineers behind it.

**How is CVE-2026-24233 attacked?** It is attacked with local access, in a software library.

**Am I affected by CVE-2026-24233?** Check whether you run Nvidia TensorRT-LLM at an affected version. Junglewise does that check for you around the clock: connect your code, cloud and domains, and it matches new vulnerabilities like this one against what you actually run and tells you when one reaches you.

## Technical details

A deserialization vulnerability (CWE-502) exists in NVIDIA TensorRT-LLM for Linux within the restricted unpickler component used during model weight loading. The flaw allows a local, unauthenticated attacker to trigger the deserialization of untrusted data by providing a specially crafted model file. Because the unpickler does not sufficiently restrict the classes it can instantiate, an attacker can achieve arbitrary code execution, escalate privileges, or access sensitive information. The vulnerability affects versions up to and including v1.3.0 rc14.

## Affected products

- NVIDIA TensorRT-LLM v1.3.0 rc14 and earlier

## Timeline

- 2026-07-14: disclosed: Initial publication of CVE-2026-24233

## References

- [https://www.cve.org/CVERecord?id=CVE-2026-24233](https://www.cve.org/CVERecord?id=CVE-2026-24233)

## Related threats

- [NVIDIA TensorRT-LLM server-side request forgery](https://junglewise.ai/threats/cve-2026-65120-nvidia-tensorrt-llm-contains-a-vulnerability-where-an-attacker)
- [NVIDIA TensorRT-LLM untrusted tokenizer loading in KvCacheAwareRouter](https://junglewise.ai/threats/cve-2026-65119-nvidia-tensorrt-llm-for-any-platform-contains-a-vulnerability-in)
- [NVIDIA TensorRT-LLM server-side request forgery](https://junglewise.ai/threats/cve-2026-24269-nvidia-tensorrt-llm-contains-a-vulnerability-where-an-attacker)
- [NVIDIA TensorRT-LLM denial of service in OpenAI-compatible API](https://junglewise.ai/threats/cve-2026-47475-nvidia-tensorrt-llm-denial-of-service-in-openai-compatible-api)
- [NVIDIA TensorRT-LLM write-what-where condition](https://junglewise.ai/threats/cve-2026-47473-nvidia-tensorrt-llm-write-what-where-condition)