CVE-2026-33298
llama.cpp has a Heap Buffer Overflow via Integer Overflow in GGUF Tensor Parsing
Description
llama.cpp is an inference of several LLM models in C/C++. Prior to b7824, an integer overflow vulnerability in the `ggml_nbytes` function allows an attacker to bypass memory validation by crafting a GGUF file with specific tensor dimensions. This causes `ggml_nbytes` to return a significantly smaller size than required (e.g., 4MB instead of Exabytes), leading to a heap-based buffer overflow when the application subsequently processes the tensor. This vulnerability allows potential Remote Code Execution (RCE) via memory corruption. b7824 contains a fix.
INFO
Published Date :
March 24, 2026, 1:17 a.m.
Last Modified :
March 24, 2026, 3:53 p.m.
Remotely Exploit :
No
Source :
[email protected]
CVSS Scores
| Score | Version | Severity | Vector | Exploitability Score | Impact Score | Source |
|---|---|---|---|---|---|---|
| CVSS 3.1 | HIGH | [email protected] |
Solution
- Update llama.cpp to b7824 or later.
- Validate GGUF file tensor dimensions.
Public PoC/Exploit Available at Github
CVE-2026-33298 has a 2 public
PoC/Exploit available at Github.
Go to the Public Exploits tab to see the list.
References to Advisories, Solutions, and Tools
Here, you will find a curated list of external links that provide in-depth
information, practical solutions, and valuable tools related to
CVE-2026-33298.
| URL | Resource |
|---|---|
| https://github.com/ggml-org/llama.cpp/releases/tag/b7824 | |
| https://github.com/ggml-org/llama.cpp/security/advisories/GHSA-96jg-mvhq-q7q7 |
CWE - Common Weakness Enumeration
While CVE identifies
specific instances of vulnerabilities, CWE categorizes the common flaws or
weaknesses that can lead to vulnerabilities. CVE-2026-33298 is
associated with the following CWEs:
Common Attack Pattern Enumeration and Classification (CAPEC)
Common Attack Pattern Enumeration and Classification
(CAPEC)
stores attack patterns, which are descriptions of the common attributes and
approaches employed by adversaries to exploit the CVE-2026-33298
weaknesses.
We scan GitHub repositories to detect new proof-of-concept exploits. Following list is a collection of public exploits and proof-of-concepts, which have been published on GitHub (sorted by the most recently updated).
Security audit documenting 221 silent int64-to-int32 truncation sites in vLLM's CUDA/C++ extensions that enable GPU buffer overflow via crafted GGUF model files.
Cathedral-Grade Security for AI Agents. 23/23 attack vectors caught. Local-first, zero API cost. MIT licensed.
Python
Results are limited to the first 15 repositories due to potential performance issues.
The following list is the news that have been mention
CVE-2026-33298 vulnerability anywhere in the article.
The following table lists the changes that have been made to the
CVE-2026-33298 vulnerability over time.
Vulnerability history details can be useful for understanding the evolution of a vulnerability, and for identifying the most recent changes that may impact the vulnerability's severity, exploitability, or other characteristics.
-
New CVE Received by [email protected]
Mar. 24, 2026
Action Type Old Value New Value Added Description llama.cpp is an inference of several LLM models in C/C++. Prior to b7824, an integer overflow vulnerability in the `ggml_nbytes` function allows an attacker to bypass memory validation by crafting a GGUF file with specific tensor dimensions. This causes `ggml_nbytes` to return a significantly smaller size than required (e.g., 4MB instead of Exabytes), leading to a heap-based buffer overflow when the application subsequently processes the tensor. This vulnerability allows potential Remote Code Execution (RCE) via memory corruption. b7824 contains a fix. Added CVSS V3.1 AV:L/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H Added CWE CWE-190 Added CWE CWE-122 Added Reference https://github.com/ggml-org/llama.cpp/releases/tag/b7824 Added Reference https://github.com/ggml-org/llama.cpp/security/advisories/GHSA-96jg-mvhq-q7q7