QW

Bonsai-27B-dspark

QwenGGUF by prism-mlUpdated 2026-07-17
Download on Hugging Face
Parameters3.6B
Context4K
Layers6
Architecturedspark
Smallest quant1.8 GB
Quick answer

Bonsai-27B-dspark needs about 4.4 GB of memory at Q1_0 (a 3.8 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 4 quants, from 1.8 GB (Q4_1) to 53.8 GB (F16).

Bonsai-27B-dspark: 27B Qwen3.6 model with vision and 1-bit quantization

Bonsai-27B-dspark is a 27B Qwen3.6 model quantized to 1-bit weights for efficient on-device inference. It supports vision and a 262,144-token context length. The model is released under Apache-2.0 and is compatible with llama.cpp, CUDA, Metal, and CPU.

Good at

  • Vision tasks
  • 1-bit efficient inference
  • Large context support

Limits

  • No high‑precision escape hatches
  • Requires custom 1-bit kernels
  • Limited to 262,144-token context

Best for

  • On‑device vision inference
  • Low‑memory inference on laptops
  • Edge deployment

Who it’s for. Developers needing a compact 27B model with vision support for edge devices. The model is suitable for applications that can run on llama.cpp or similar frameworks with custom low‑bit kernels.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_11.8 GB 2.3 GB
Fits Get file →
Q1_03.8 GB 4.4 GB
Fits Get file →
BF167.3 GB 7.8 GB
Tight Get file →
F1653.8 GB 54.4 GB
Too big Get file →

For 8 GB VRAM, use Q4_1 (1.79 GB file, ~2.4 GB memory). For 4 GB VRAM, use Q1_0 (3.8 GB file, ~4.4 GB memory). The BF16 and F16 variants require larger memory and are not recommended for low‑budget GPUs.

What GPU do you need for Bonsai-27B-dspark?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is Q4_1 (1.8 GB file, about 2.4 GB in use).
  • 6 GB VRAM: the best quant that fits is Q1_0 (3.8 GB file, about 4.4 GB in use).
  • 8 GB VRAM: the best quant that fits is BF16 (7.3 GB file, about 7.9 GB in use).
  • 12 GB VRAM: the best quant that fits is BF16 (7.3 GB file, about 7.9 GB in use).
  • 16 GB VRAM: the best quant that fits is BF16 (7.3 GB file, about 7.9 GB in use).
  • 24 GB VRAM: the best quant that fits is BF16 (7.3 GB file, about 7.9 GB in use).
  • 32 GB VRAM: the best quant that fits is BF16 (7.3 GB file, about 7.9 GB in use).
  • 48 GB VRAM: the best quant that fits is BF16 (7.3 GB file, about 7.9 GB in use).

Which Bonsai-27B-dspark quant should you choose?

Q1_0 (3.8 GB) is the middle option in this repository. Lower bit counts mean smaller files and more quality loss.

If memory is tight, the smallest file is Q4_1 at 1.8 GB, with the quality trade-offs that come with fewer bits. The usual sweet spot between file size and quality, and the quant most people start with.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturedspark
Parameters3.6B
Layers6
Trained context4,096 tokens
KV cache (fp16)12 KB per token

Source

Quantized byprism-ml
FamilyQwen
Licenseapache-2.0
Downloads (HF)283,794
Last updated2026-07-17

Frequently asked questions about Bonsai-27B-dspark

How much VRAM does Bonsai-27B-dspark need?

About 4.4 GB at Q1_0 with an 8K context window. That is the 3.8 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Bonsai-27B-dspark on an 8 GB GPU?

Yes, with BF16: the 7.3 GB file needs about 7.9 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Bonsai-27B-dspark GGUF?

It depends on the quant: Q4_1 is 1.8 GB, Q1_0 is 3.8 GB, BF16 is 7.3 GB, F16 is 53.8 GB.

What context length does Bonsai-27B-dspark support?

The GGUF header lists a trained context length of 4,096 tokens (4K). The KV cache grows with the context you use: about 0.1 GB at 8K and 0.1 GB at 4K.

Where can I download Bonsai-27B-dspark GGUF?

The files are on Hugging Face in the prism-ml/Bonsai-27B-gguf repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Bonsai-27B-dspark use?

Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run Bonsai-27B-dspark on 8 GB VRAM?

Yes, the Q4_1 quantized file is 1.79 GB and requires about 2.4 GB of VRAM at 8 k context.

Does this model support vision?

Yes, Bonsai-27B-dspark includes a vision encoder and is detected to have vision abilities.

What license does it use?

The model is released under the Apache-2.0 license.

Other GGUFs of Qwen3.6-27B

Similar models

Browse all →

More from prism-ml

All by prism-ml →

More Qwen models

All Qwen →

Other small models

All small models →