1bit engine: local LLMs on Strix Halo
One binary to rule them all: model-agnostic local LLM inference on the AMD Ryzen AI MAX+ 395 (Strix Halo), any model on the Radeon 8060S iGPU and the XDNA 2 NPU, inside AMD's Lemonade Server. Boot a stick and it serves.
latest · 2026-10-30This week in the engine. The engine is pinned to this week's Lemonade release, the release candidate's checks are green, and two measurements this week say plainly where two slow things stand.read the post →
site index
Run local LLMs on Strix Halo: guides, docs and measured results
Engine →
What the engine is, what runs where, and who it is built on.
engineStrix Halo guide →Run local LLMs on AMD Strix Halo (Ryzen AI MAX+ 395) on Linux: the Radeon 8060S iGPU through HRX, the XDNA 2 NPU,…
strix-haloFAQ →Answers about local LLMs with 1bit engine: how much memory a model needs, which fit in 128 GB, Linux vs Windows, the…
faqDocs →Every device, every component, and how to run them.
docsBlog →The build log: what moved over, what got faster, what broke.
blogBenchmarks →Measured, not projected. Every number with how it was taken.
wiki