The ggml‑org release on Oct 8, 2026 patches a CUDA memory issue and provides pre‑built binaries for macOS, iOS and several Linux platforms.
NoteBy OossaPublished by Oossa: 1 min read
The open‑source llama.cpp library released version b11511 on Oct 8, 2026. It fixes a CUDA out‑of‑bounds read bug in the MMQ implementation. The update also adds ready‑to‑run binaries for macOS (Apple Silicon and Intel), iOS, and Ubuntu on x64, arm64 and s390x, plus Vulkan and CUDA‑12.8 builds.
Developers can now run llama.cpp on more hardware without crashes, especially those using NVIDIA GPUs.
Models
The company’s new three‑agent Claude‑based reviewer found 73.43% of core‑claim errors, far above the previous best of 14.81%.
Models
Mistral's new trillion‑parameter model scores 38 on an independent index, but costs more per task than competing open models.
Models
OpenAI reported two influence campaigns that used AI‑generated content to plant false stories in real media and blocked the involved accounts.
Oossa · Newsletter
The week in AI, explained
Every Monday: the stories worth knowing, in plain language. Free, no spam.
