NVIDIA Megatron-LM Cache Hash Collision Vulnerability (CVE-2026-65104)
CVE-2026-65104 Published on October 13, 2026
NVIDIA Megatron-LM for all platforms contains a vulnerability in inference, where an attacker could cause hash collisions in the cache used for inference requests. A successful exploit of this vulnerability might lead to data tampering and partial information disclosure.
Vulnerability Analysis
CVE-2026-65104 is exploitable with network access, and requires small amount of user privileges. This vulnerability is consided to have a high level of attack complexity. The potential impact of an exploit of this vulnerability is considered to have a small impact on confidentiality, a high impact on integrity, and no impact on availability.
Weakness Type
Reversible One-Way Hash
The product uses a hashing algorithm that produces a hash value that can be used to determine the original input, or to find an input that can produce the same hash, more efficiently than brute force techniques. This weakness is especially dangerous when the hash is used in security algorithms that require the one-way property to hold. For example, if an authentication system takes an incoming password and generates a hash, then compares the hash to another hash that it has stored in its authentication database, then the ability to create a collision could allow an attacker to provide an alternate password that produces the same target hash, bypassing authentication.
Products Associated with CVE-2026-65104
Want to know whenever a new CVE is published for NVIDIA Megatron Lm? stack.watch will email you.
