Ternary Bonsai 8B needs about 3.9 GB of memory at Q2_0 (a 2.2 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 4 quants, from 2.2 GB to 16.4 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 4.8 GB.
Ternary Bonsai 8B – a 1.58‑bit Q2_0 GGUF model for efficient inference
Ternary Bonsai 8B is an 8‑billion‑parameter Qwen3 model quantized to Q2_0 for llama.cpp. It supports a 65,536‑token context window and is released under Apache‑2.0. The model is available in multiple GGUF files with different sizes and memory estimates.
Good at
- Fast inference on CPU and Metal
- Low memory footprint with Q2_0
- Large 65k‑token context
Limits
- Requires custom llama.cpp fork for Q2_0
- No vision or multimodal support
- License is Apache‑2.0
Best for
- Running on edge devices with limited VRAM
- Serving low‑latency chat applications
- Fine‑tuning with small GPU budgets
Who it’s for. Developers and researchers needing a compact 8B model that fits in modest hardware. The model is suitable for local deployment on Apple Silicon or CPUs with NEON support, provided the custom llama.cpp fork is used.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 2.2 GB | 3.9 GB | Fits | Get file → | |
| Q2_0 | 2.2 GB | 3.9 GB | Fits | Get file → | |
| Q2_0 | 2.3 GB | 4.0 GB | Fits | Get file → | |
| F16 | 16.4 GB | 18.1 GB | Too big | Get file → |
For 8 GB VRAM, use the Q2_0 2.18 GB file; it needs about 3.9 GB of RAM at 8 k context. If more memory is available, the 2.31 GB Q2_0 file uses roughly 4 GB. The FP16 16.38 GB file requires about 18 GB of RAM and is only useful on high‑end GPUs.
What GPU do you need for Ternary Bonsai 8B?
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 Q2_0 (2.3 GB file, about 4.0 GB in use).
- 6 GB VRAM: the best quant that fits is Q2_0 (2.3 GB file, about 4.0 GB in use).
- 8 GB VRAM: the best quant that fits is Q2_0 (2.3 GB file, about 4.0 GB in use).
- 12 GB VRAM: the best quant that fits is Q2_0 (2.3 GB file, about 4.0 GB in use).
- 16 GB VRAM: the best quant that fits is Q2_0 (2.3 GB file, about 4.0 GB in use).
- 24 GB VRAM: the best quant that fits is F16 (16.4 GB file, about 18.1 GB in use).
Which Ternary Bonsai 8B quant should you choose?
Q2_0 (2.2 GB) is the middle option in this repository. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
If memory is tight, the smallest file is the first one in the table at 2.2 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Ternary Bonsai 8B
How much VRAM does Ternary Bonsai 8B need?
About 3.9 GB at Q2_0 with an 8K context window. That is the 2.2 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 Ternary Bonsai 8B on an 8 GB GPU?
Yes, with Q2_0: the 2.3 GB file needs about 4.0 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Ternary Bonsai 8B GGUF?
It depends on the quant: file is 2.2 GB, Q2_0 is 2.2 GB, Q2_0 is 2.3 GB, F16 is 16.4 GB.
What context length does Ternary Bonsai 8B support?
The GGUF header lists a trained context length of 65,536 tokens (64K). The KV cache grows with the context you use: about 1.2 GB at 8K and 4.8 GB at 32K.
Where can I download Ternary Bonsai 8B GGUF?
The files are on Hugging Face in the prism-ml/Ternary-Bonsai-8B-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 Ternary Bonsai 8B 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 Ternary Bonsai 8B on 8 GB VRAM?
Yes, the Q2_0 2.18 GB file uses about 3.9 GB of RAM at 8 k context, fitting comfortably on 8 GB VRAM setups.
Do I need a special version of llama.cpp?
Q2_0 is not in mainline llama.cpp; use the PrismML-Eng/llama.cpp fork on the prism branch for CPU and Metal support.
What context length does the model support?
The model was trained with a 65,536‑token context window and can be used up to that length.