The open‑source llama.cpp library released version b11493 on Oct 8 2026, adding a new SYCL‑based grouped MoE XMX GEMM operation and binaries for many platforms.
NoteBy OossaPublished by Oossa: 1 min read
The ggml‑org team pushed llama.cpp version b11493 on Oct 8 2026. The update adds a SYCL implementation of grouped Mixture‑of‑Experts (MoE) XMX GEMM, a matrix multiply used in large language models. The release also bundles pre‑built binaries for macOS, iOS, Ubuntu (CPU, Vulkan and CUDA).
Developers can now run MoE‑based models on SYCL‑compatible GPUs, expanding hardware options for open‑source LLM workloads.
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.
