BLAST RADIUS

Open-weight & owned stack · Quantization efficiency

llama.cpp Adds Arbitrary Metal HC Support

Sep 21, 2026, first seen via Training & inference tooling releases

llama.cpp releases b11067 through b11069 added arbitrary hc support to Metal dsv4_hc_pre kernels.

Read the original at github.comOpens the publisher's site in a new tab

Also covering this

b11068github.com, Sep 21b11067github.com, Sep 21

More in Open-weight & owned stack

Z.ai Releases Open-Weight GLM-5.2 ModelSep 21Nvidia Promotes Nemotron Open Weights UAESep 21Developers Criticize Qwen Image 2.1 LicenseSep 21llama.cpp Expands Hexagon Memory SupportSep 21