# CVE-2026-65180: NVIDIA Megatron-LM remote code execution in checkpoint loading

Source: https://junglewise.ai/threats/cve-2026-65180-nvidia-megatron-lm-contains-a-vulnerability-in-checkpoint-loading  
Publisher: Junglewise Threat Intelligence, https://junglewise.ai/threats  
Published: 2026-10-13. Updated: 2026-10-16  
License: CC BY 4.0 with a link back

CVE-2026-65180 · Severity: high · CVSS 7.8 · Published 2026-10-13

NVIDIA Megatron-LM is a large language model training framework used for deep learning research and production deployments. An attacker can achieve remote code execution by tricking a user into loading a maliciously crafted checkpoint file, potentially leading to system compromise, privilege escalation, and data theft.

Technologies: [Nvidia Megatron-LM](https://junglewise.ai/threats/technologies/megatron-lm). Vendors: [Nvidia](https://junglewise.ai/threats/vendors/nvidia).

## Quick answers

**Is CVE-2026-65180 exploited in the wild?** No. Junglewise Threat Intelligence has no report of CVE-2026-65180 being exploited in the wild.

**How severe is CVE-2026-65180?** High severity, with a CVSS score of 7.8, published 13 October 2026.

**Which products and versions are affected?** Nvidia Megatron-LM <UNKNOWN>.

**Is there a fix for CVE-2026-65180?** 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-2026-65180 attacked?** It is attacked with local access, in a software library.

**Am I affected by CVE-2026-65180?** Check whether you run Nvidia Megatron-LM 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

The vulnerability exists in the checkpoint loading mechanism of Megatron-LM, where deserialization of malicious checkpoint files can lead to arbitrary code execution. An attacker must convince a user to load a crafted file, making this a user-interaction attack vector. A successful exploit grants full code execution and privilege escalation capabilities.

## Affected products

- NVIDIA Megatron-LM <UNKNOWN>

## Timeline

- 2026-10-13: disclosed

## References

- [https://github.com/NVIDIA/product-security/tree/main/2026/5881](https://github.com/NVIDIA/product-security/tree/main/2026/5881)
- [https://www.cve.org/CVERecord?id=CVE-2026-65180](https://www.cve.org/CVERecord?id=CVE-2026-65180)

## Related threats

- [NVIDIA Megatron-LM remote code execution in checkpoint loading](https://junglewise.ai/threats/cve-2026-65110-nvidia-megatron-lm-contains-a-vulnerability-in-checkpoint-loading)
- [NVIDIA Megatron-LM hash collision in inference cache](https://junglewise.ai/threats/cve-2026-65104-nvidia-megatron-lm-for-all-platforms-contains-a-vulnerability-in)