CVE-2026-27940
High · 7.8CVSS
7.8
EPSS
0.2%
Published
Mar 12, 2026
CISA due date
—
The record
llama.cpp is an inference of several LLM models in C/C++. Prior to b8146, the gguf_init_from_file_impl() in gguf.cpp is vulnerable to an Integer overflow, leading to an undersized heap allocation. Using the subsequent fread() writes 528+ bytes of attacker-controlled data past the buffer boundary. This is a bypass of a similar bug in the same file - CVE-2025-53630, but the fix overlooked some areas. This vulnerability is fixed in b8146.
Affected packages
| Ecosystem | Package | Introduced | Fixed |
|---|
Related vulnerabilities
- CVE-2026-112403 Pillow is a Python imaging library. CWE-122
- CVE-2026-112393 Pillow is a Python imaging library. CWE-122
- CVE-2026-112395 Pillow is a Python imaging library. CWE-122
- CVE-2026-103658 GNU Aspell 0.60.8.2 (first vulnerable version is not known) contains a heap-based buffer… CWE-122
- CVE-2026-107161 A heap-based buffer overflow flaw was found in Cyrus SASL. CWE-122
- CVE-2026-106574 ImageMagick is free and open-source software used for editing and manipulating digital… CWE-122
- CVE-2026-106564 ImageMagick is free and open-source software used for editing and manipulating digital… CWE-122
- CVE-2026-46570 In NTFS-3G before 2026.7.7, a heap buffer overflow exists in ntfs_index_walk_down() in… CWE-122