CVE-2026-65104: NVIDIA Megatron-LM hash collision in inference cache

CVE-2026-65104 · Severity: high · CVSS 7.1 · Published 2026-10-13

NVIDIA Megatron-LM is a framework for training and deploying large language models. An attacker could trigger hash collisions in the inference cache, allowing them to tamper with cached data or access information they should not see, impacting data integrity and confidentiality.

Technologies: Nvidia Megatron-LM. Vendors: Nvidia.

Quick answers

Is CVE-2026-65104 exploited in the wild?
No. Junglewise Threat Intelligence has no report of CVE-2026-65104 being exploited in the wild.
How severe is CVE-2026-65104?
High severity, with a CVSS score of 7.1, published 13 October 2026.
Which products and versions are affected?
Nvidia Megatron-LM all platforms.
Is there a fix for CVE-2026-65104?
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-2026-65104 attacked?
It is attacked over the network, in a software library.
Am I affected by CVE-2026-65104?
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

A hash collision vulnerability exists in the inference cache mechanism of Megatron-LM. An attacker with network access to inference requests can cause deliberate collisions in the cache, leading to data tampering and information disclosure. The vulnerability affects all platforms and requires no authentication.

Affected products

  • NVIDIA Megatron-LM all platforms

Timeline

  • 2026-10-13: disclosed

References

This page as Markdown. Free to reuse under CC BY 4.0 with a link back.