The open‑source LLM runtime adds an L2‑friendly workgroup scheduling tweak and ships updated binaries for macOS, iOS and Ubuntu.
NoteBy OossaPublished by Oossa: 1 min read
The ggml‑org team published llama.cpp version b11640. The update changes the Vulkan backend to use L2‑friendly workgroup scheduling for matrix multiplication. It also provides fresh pre‑built binaries for Apple Silicon, Intel Macs, iOS, and several Ubuntu builds (CPU, Vulkan and CUDA 12.8). Users can download the packages from the GitHub release page.
The change can make LLaMA models run faster on GPUs that support Vulkan, helping developers on Linux and Android devices.
Models
The new Kandinsky S2V 1.0 Lite (2B) and Pro (21B) models turn speech and a single photo into 25 FPS talking‑head video, and the code is free under MIT.
Models
The new model scores 86.9 on Artificial Analysis' Speech-to-Speech Index, beating OpenAI and Google, and is now live in the Voice Agent API.
Models
YouTube Music will mark tracks that use generative AI with “Fully Gen AI,” “Partly Gen AI,” or “No Gen AI” labels, starting Oct 14 2026.
Oossa · Newsletter
The week in AI, explained
Every Monday: the stories worth knowing, in plain language. Free, no spam.
