Prism ML Website  |  Demo & Examples  |  Discord

Bonsai 2 Ternary 2B: GGUF

A 2B-class vision-language model with ternary transformer weights, for llama.cpp (CUDA, Metal, CPU).

~0.52 GB language model (PQ2_0), down from ~3.8 GB at FP16

Highlights

• Ternary language weights across embeddings, attention projections, MLP projections and the (tied) LM head, stored with a group-wise FP16 scale.
• Three GGUF packings of the same weights, consumed directly by custom kernels and never expanded back to FP16: PTQ1_0 (dense trits, 0.43 GB), PQ2_0 (2-bit slots, 0.52 GB) and Q2_0 (0.55 GB).
• Vision input through a separate mmproj pack (Q8_0, or BF16 for reference precision), loaded only when an image arrives.
• 262K-token context, inherited from the Qwen3.5-2B hybrid-attention backbone (18 linear-attention and 6 full-attention layers).

Model Overview

Files

Bits/weight is file size over the ~1.88B language-model parameters, so it includes the F32 tensors and metadata. The three language-model packs hold the same ternary weights; they differ only in storage layout and kernel speed.

Quickstart

PrismML-Eng/Bonsai-demo is the source of truth for running these models. It has the tested setup for every backend and pins a known-good binary release.

Use the PrismML llama.cpp build

The ternary kernels live in the PrismML-Eng/llama.cpp fork (branch prism). It runs every file here. Upstream llama.cpp runs only Q2_0 (recent master, which includes the Hadamard activation runtime); it rejects PQ2_0 and PTQ1_0 as unknown types.

# prebuilt: pick the archive for your platform
# https://github.com/PrismML-Eng/llama.cpp/releases/latest
tar -xzf llama-<tag>-bin-<platform>.tar.gz -C bin --strip-components=1

# or build it
git clone https://github.com/PrismML-Eng/llama.cpp && cd llama.cpp
cmake -B build -DGGML_CUDA=ON && cmake --build build -j   # drop -DGGML_CUDA=ON on macOS; Metal is the default

Text

./bin/llama-cli -m Ternary-Bonsai-2-2B-PQ2_0.gguf -ngl 99 -fa on \
  -p "Explain in three sentences why the sky is blue."

Image + text

./bin/llama-mtmd-cli -m Ternary-Bonsai-2-2B-PQ2_0.gguf \
  --mmproj Ternary-Bonsai-2-2B-mmproj-Q8_0.gguf -ngl 99 \
  --image photo.jpg -p "Describe this image."

Vision projector: use the Q8_0 mmproj. With the current fork build, the BF16 mmproj can crash llama-server when several image requests run at once; Q8_0 is not affected and quality is nearly the same.

Small text in images: for small or dense text, raising the image token budget with --image-min-tokens 1024 may help.

OpenAI-compatible server

./bin/llama-server -m Ternary-Bonsai-2-2B-PQ2_0.gguf \
  --mmproj Ternary-Bonsai-2-2B-mmproj-Q8_0.gguf -ngl 99 -fa on -c 32768 --port 8080

curl http://localhost:8080/v1/chat/completions -H "Content-Type: application/json" -d '{
  "messages": [{"role": "user", "content": [
    {"type": "text", "text": "What is in this image?"},
    {"type": "image_url", "image_url": {"url": "https://upload.wikimedia.org/wikipedia/commons/4/47/PNG_transparency_demonstration_1.png"}}
  ]}],
  "chat_template_kwargs": {"enable_thinking": false}
}'

Best Practices

Generation Parameters

We recommend these sampling settings. They are the same for every packing and for MLX:

• Thinking mode (default): temperature=1.0, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=1.0, repetition_penalty=1.0
• Non-thinking mode: temperature=0.7, top_p=0.80, top_k=20, min_p=0.0, presence_penalty=1.5, repetition_penalty=1.0

Apply the penalties over the last 64 tokens (llama.cpp's default window). Presence penalty has no GGUF metadata key, so pass it on the command line or per request.

# thinking on
./bin/llama-server -m Ternary-Bonsai-2-2B-PQ2_0.gguf -ngl 99 -fa on \
  --temp 1.0 --top-p 0.95 --top-k 20 --min-p 0 --presence-penalty 1.0 --repeat-penalty 1.0 --repeat-last-n 64

# thinking off: start the server with these values, and send
# "chat_template_kwargs": {"enable_thinking": false} with each request
./bin/llama-server -m Ternary-Bonsai-2-2B-PQ2_0.gguf -ngl 99 -fa on \
  --temp 0.7 --top-p 0.8 --top-k 20 --min-p 0 --presence-penalty 1.5 --repeat-penalty 1.0 --repeat-last-n 64

Avoid greedy decoding (temperature=0) with thinking on; it can get stuck reasoning on some prompts.

Output Length

Give thinking mode room to finish. Even simple prompts can use 3,000+ thinking tokens, so allow at least 8,192 output tokens with thinking on. For short answers, turn thinking off.

System Prompt

You can use a simple system prompt such as:

You are a helpful assistant

MLX

On Apple Silicon you can also use the MLX build of the same weights, Ternary-Bonsai-2-2B-mlx-2bit. It loads natively in mlx-vlm 0.7.8 and newer (the released mlx-lm cannot load it yet). Pass the values above explicitly: mlx-vlm's Python generate() is greedy by default. Its model card has the commands.

Choosing a Packing

• PQ2_0: the default; cheaper to unpack, and the pack measured on Apple Silicon.
• PTQ1_0: the smallest file; use it where memory is tightest.
• Q2_0: an alternative layout, and the one format upstream llama.cpp can also run.

Speculative Decoding

The 2B ships without a draft model. At this size a decode step is already cheap, and in our tests speculative decoding with a draft head was slower than plain decoding on everyday prompts, so plain decoding is the recommended setup.

License

Apache 2.0, following the base model.