CVE-2025-49847
CVE-2025-49847 · llama.cpp Vulnerable to Buffer Overflow via Malicious GGUF Model
This page contains facts from public vulnerability sources. Check the original advisory before updating.
llama.cpp is an inference of several LLM models in C/C++. Prior to version b5662, an attacker‐supplied GGUF model vocabulary can trigger a buffer overflow in llama.cpp’s vocabulary‐loading code. Specifically, the helper _try_copy in llama.cpp/src/vocab.cpp: llama_vocab::impl::token_to_piece() casts a very large size_t token length into an int32_t, causing the length check (if (length < (int32_t)size)) to be bypassed. As a result, memcpy is still called with that oversized size, letting a malicious model overwrite memory beyond the intended buffer. This can lead to arbitrary memory corruption and potential code execution. This issue has been patched in version b5662.
CVSS 8.8 · High · EPSS 0.5% · Fix available
- Published
- Updated
Stack: system
Affected products and versions
llama.cpp
- Versions
- < b5662: Affected
No fix published
llama.cpp Debian:14
- Versions
- 5151+dfsg-1~exp2
- 5151+dfsg-1~exp3
- 5318+dfsg-1
- 5318+dfsg-2
- Affected ranges
- ECOSYSTEM: < 5713+dfsg-1
Update to 5713+dfsg-1
llama.cpp Ubuntu:25.10
- Versions
- 5318+dfsg-1
- 5318+dfsg-2
- 5713+dfsg-1
- 5760+dfsg-1
- 5760+dfsg-3
- 5760+dfsg-4
- 5882+dfsg-2
- Affected ranges
- ECOSYSTEM: (−∞, +∞)
No fix published
llama.cpp Ubuntu:26.10
- Versions
- 8681+dfsg-1
- 9721+dfsg-1ubuntu1
- 10438+dfsg-1
- 10438+dfsg-1ubuntu1
- Affected ranges
- ECOSYSTEM: (−∞, +∞)
No fix published
llama.cpp Ubuntu:26.04:LTS
- Versions
- 5882+dfsg-2
- 8064+dfsg-1ubuntu1
- 8681+dfsg-1
- Affected ranges
- ECOSYSTEM: (−∞, +∞)
No fix published