The open‑source LLM runtime releases version b12329 with CUDA optimizations and pre‑built binaries for macOS, iOS and Linux.
NoteBy OossaPublished by Oossa: 1 min read
The ggml‑org team released llama.cpp version b12329. It adds a CUDA optimisation that fuses weight‑scale handling into matrix‑multiply writes, speeding up nvfp4 checkpoints. The update also includes new pre‑built binaries for Apple Silicon, Intel macOS, iOS and several Linux flavours. The changes are available for download from the GitHub release page.
Developers can run Llama models faster on NVIDIA GPUs and get ready‑to‑run binaries for more platforms.
Models
The new Gemini 4 Hydrogen model beats the larger Argon on speed, cost and several benchmarks, and is now available to developers and consumers worldwide.
Models
The new Gemini 4 Hydrogen model is now available across Google AI products, offering faster, cheaper and more reliable responses than the previous 3.8 Flash.
Models
Google says the new Gemini 4 Hydrogen model is now live for all Search users, promising deeper understanding, broader link results and richer visual answers.
Oossa · Newsletter
The week in AI, explained
Every Monday: the stories worth knowing, in plain language. Free, no spam.
