Ternary Bonsai 2

Ternary Bonsai 2 (27B) released under 6GB

Ternary Bonsai 2, derived from Qwen3.8-27B, keeps the 27B hybrid‑attention architecture but uses ternary weights to shrink the model to under 6GB; the model card claims it is 9× smaller than FP16 while retaining 98.2% of its intelligence, and it’s available on Hugging Face with a WebGPU demo.