Bonsai 1.7B is an open model by Prism ML: built from Qwen3-1.7B and trained end-to-end for 1-bit weights — 0.24 GB instead of 3.4 GB in FP16, running entirely in your browser on WebGPU. This is the free-chat variant: talk to the model directly and see what a 1-bit 1.7B can (and can’t) do. Nothing you type leaves your device. A sysf.io demonstration of sovereign, on-device inference.
1.7B PARAMS · 1-BIT Q1_0 (≈1.125 BPW) · 0.24 GB · 32K CONTEXT · APACHE-2.0
MODEL © PRISM ML (APACHE-2.0) · ENGINE: WEBML-COMMUNITY (MIT), PORTED TO QWEN3 BY SYSF.IO · TRIAGE VARIANT
Enter a Hugging Face access token with read access to prism-ml/Bonsai-1.7B-gguf. It stays in this browser — requests go only to huggingface.co.
You’re chatting with a 1-bit, 1.7-billion-parameter model running on your GPU. Expect quick, concise answers — and the occasional small-model stumble. Nothing leaves your device.
Kernels are the low-level GPU programs that do the model's actual math — the 1-bit matrix multiplies, attention and normalizations behind every token. Each one is a WGSL compute shader, compiled by your browser for this GPU.