CVE-2026-70640 · UNKNOWN
Summary: llama.cpp b1886–b7445 Race Condition Use-After-Free via llama-android.cpp
Details
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_1context() lack synchronization, allowing Thread A to operate on freed memory while Thread B concurrently frees the llama_context. Attackers can exploit this by performing heap spray with attacker-controlled data containing a fake vtable to hijack the vtable pointer at offset +0x30, causing llama_batch_allocr::clear() to dereference arbitrary memory and achieve remote code execution.
Severity
UNKNOWN — vector CVSS:4.0/AV:L/AC:H/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N
Affected packages
- GIT github.com/ggml-org/llama.cpp [862f5e41ab1fdf12d6f59455aad3f5dd8258f805 – 5c0d18881e0e9794c96b2602736b758bac9d9388]
References
- https://www.cyera.com/research/breaking-local-ai-runtimes-10-vulnerabilities-in-the-engine-behind-your-open-source-models
- https://www.youtube.com/watch?v=_FKX5hDwgVE&t=1079s
- https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/70xxx/CVE-2026-70640.json
- https://github.com/ggml-org/llama.cpp/releases/tag/b7446
- https://nvd.nist.gov/vuln/detail/CVE-2026-70640
- https://www.vulncheck.com/advisories/llama-cpp-b1886-b7445-race-condition-use-after-free-via-llama-android-cpp
- https://github.com/ggml-org/llama.cpp/commit/5c0d18881e0e9794c96b2602736b758bac9d9388
- https://github.com/Vladimir-tokarev-cyera/llama-cpp-security-patches