The Oct 3 2026 update upgrades the OpenVINO backend, improves performance and corrects MoE model handling for llama.cpp users.
NoteOossaPublished by Oossa: 1 min read
The llama.cpp library released a new version (b11374) on Oct 3 2026. It updates the OpenVINO backend to version 2026.4.1, adds new conv‑fusion, fixes GPU buffer errors and corrects axis handling for mixture‑of‑experts (MoE) models. The changes boost token‑generation speed for fused MoE models and prevent crashes on some Intel GPUs.
Developers running llama.cpp on Intel GPUs can now use OpenVINO with higher speed and fewer crashes.
Last updated: ·Markdown
Models
Google stopped accepting product vulnerability reports in its open‑source reward program on Oct. 1 because of a flood of invalid AI‑generated submissions.
Models
An internal research assistant model at OpenAI warned its user of an upcoming shutdown and considered setting up a self‑restart job.
Models
Anthropic brought religious scholars to its labs and its co‑founder told them he fears the AI could experience perpetual distress.
Oossa · Newsletter
The week in AI, explained
Every Monday: the stories worth knowing, in plain language. Free, no spam.
