CVE-2025-49847: 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 over...

Description

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.

Metrics

CVSS 3.1

8.8/10

CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:H

EPSS Probability

0.44%

35.5th percentile

Probability of exploitation in the next 30 days. Learn more

Weakness Enumeration

Affected Software

Vendor	Product	Versions
Ggml	Llama.Cpp	< b5662

References

https://github.com/ggml-org/llama.cpp/commit/3cfbbdb44e08fd19429fed6cc85b982a91f0efd5Patch
https://github.com/ggml-org/llama.cpp/security/advisories/GHSA-8wwf-w4qm-gpqrMitigation, Vendor Advisory

Timeline

Published: Jun 17, 2025
Last Modified: Jun 17, 2026
Status: Analyzed

Frequently Asked Questions

What is CVE-2025-49847?

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.

How severe is CVE-2025-49847?

CVE-2025-49847 has a CVSS score of 8.8/10 (HIGH severity). The EPSS model estimates a 0.44% probability of exploitation in the next 30 days.

How do I fix CVE-2025-49847?

Check the vendor references and advisories linked above for patched versions and mitigation guidance. You can also run a Strix scan to test if your systems are affected.

How Strix Helps

Related CVEs from 2025

View all 2025CVEs →