Qwen 3.8 27b is what I run locally and what I used in this tutorial video. Now with this compressed model opens up the ability to run effectively the same thing on a Mac with 16GB of RAM. Wild!
Run Bonsai 27B locally on a 16 GB Mac đĨˇ
Bonsai 2 27B is @PrismML's ternary build of Qwen3.8 27B that keeps 98.2% of FP16 quality in 7 GB, it made a voxel Japanese pagoda in one prompt with Three.js!
Run AI models locally -