News
Bonsai 2 27B: 5.9GB Model Keeps 98% of Qwen3.8’s Brain
PrismML's Bonsai 2 27B fits Qwen3.8 27B in just 5.9GB via ternary quantization, retaining 98.2% ...
