Ternary-Bonsai-2-27B-mlx-2bit 2-bit quantized model released
September 17, 2026
Prism-ML released a 2-bit quantized version of the Ternary-Bonsai-2-27B model optimized for MLX. The extreme quantization targets high-efficiency inference on Apple silicon hardware.
HOW THIS AFFECTS YOU
●
builderYou can test ultra-low precision inference for 27B parameter models on local Mac hardware.