ggml-org vulnerabilities

By year

Bars are flaws published each year; the red part is the share attackers have used.

Latest high and critical flaws

• CVE-2026-118375 llama.cpp b8841 before b11865 contains an out-of-bounds write vulnerability in the ggml-rpc RDMA transport rdma_recv() function that allows unauthenticated peerCritical 9.2 · Oct 27, 2026
• CVE-2026-116816 llama.cpp b5899 before b10261 contains a heap out-of-bounds write vulnerability in llm_tokenizer_plamo2::build that allows attackers to corrupt heap memory via High 8.5 · Oct 26, 2026
• CVE-2024-58400 llama.cpp before b3701 contains an out-of-bounds read vulnerability in llama_decode_internal and llama_encode_internal that allows remote unauthenticated attackHigh 8.8 · Oct 26, 2026
• CVE-2026-107183 llama.cpp before b11393 contains a use-after-free and double free vulnerability in common_chat_peg_mapper::map that allows unauthenticated remote attackers to cCritical 9.2 · Oct 7, 2026
• CVE-2026-70640 llama.cpp builds b1886 through b7445 contain a race condition use-after-free vulnerability in the LLaMA-Android JNI wrapper where bench_1model() and free_1conteHigh 7.3 · Aug 6, 2026
• CVE-2026-70638 llama.cpp builds b1886 through b7445 contain an integer overflow vulnerability in the LLaMA-Android JNI wrapper where the new_1batch() function multiplies sizeoHigh 8.5 · Aug 6, 2026
• CVE-2026-43632 llama.cpp builds b7492 through the latest b9060 contains a use-after-free vulnerability in llama-server affecting six tokenization endpoints (/tokenize, /detokeCritical 9.2 · Aug 6, 2026
• CVE-2026-43631 llama.cpp builds b7492 through the latest b9060 contains a use-after-free vulnerability in the vocab pointer of llama-server when the --sleep-idle-seconds featuCritical 9.2 · Aug 6, 2026
• CVE-2026-43629 llama.cpp builds b4882 through b9058 contain a heap buffer overflow vulnerability in the KV cache state restore path where the state_read_data() function computCritical 9.2 · Aug 6, 2026
• CVE-2026-43628 llama.cpp builds b3978 through b9058 contain an integer underflow and out-of-bounds read vulnerability in the DRY sampler that allows unauthenticated attackers High 8.5 · Aug 6, 2026
• CVE-2026-43627 llama.cpp builds b1283 through b9058 contain an integer overflow vulnerability in the llama_batch_init() function where unchecked multiplications in malloc() caHigh 8.5 · Aug 6, 2026
• CVE-2026-43622 llama.cpp builds b1886 through b7445 contain a double free vulnerability in the LLaMA-Android JNI wrapper where new_1batch() allocates memory using malloc() whiHigh 8.5 · Aug 6, 2026
• CVE-2026-34159 llama.cpp is an inference of several LLM models in C/C++.Critical 9.8 · Apr 1, 2026
• CVE-2026-33298 llama.cpp is an inference of several LLM models in C/C++.High 7.8 · Mar 23, 2026
• CVE-2026-27940 llama.cpp is an inference of several LLM models in C/C++.High 7.8 · Mar 12, 2026
• CVE-2026-21869 llama.cpp is an inference of several LLM models in C/C++.High 8.8 · Jan 7, 2026
• CVE-2025-53630 llama.cpp is an inference of several LLM models in C/C++.High 8.9 · Jul 10, 2025
• CVE-2025-52566 llama.cpp is an inference of several LLM models in C/C++.High 8.6 · Jun 24, 2025
• CVE-2025-49847 llama.cpp is an inference of several LLM models in C/C++.High 8.8 · Jun 17, 2025