Instructions to use prism-ml/Ternary-Bonsai-27B-gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Inference
HuggingChat
Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use prism-ml/Ternary-Bonsai-27B-gguf with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf prism-ml/Ternary-Bonsai-27B-gguf:F16
# Run inference directly in the terminal:
llama cli -hf prism-ml/Ternary-Bonsai-27B-gguf:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf prism-ml/Ternary-Bonsai-27B-gguf:F16
# Run inference directly in the terminal:
llama cli -hf prism-ml/Ternary-Bonsai-27B-gguf:F16

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf prism-ml/Ternary-Bonsai-27B-gguf:F16
# Run inference directly in the terminal:
./llama-cli -hf prism-ml/Ternary-Bonsai-27B-gguf:F16

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf prism-ml/Ternary-Bonsai-27B-gguf:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf prism-ml/Ternary-Bonsai-27B-gguf:F16

Use Docker

docker model run hf.co/prism-ml/Ternary-Bonsai-27B-gguf:F16

LM Studio
Jan

vLLM

How to use prism-ml/Ternary-Bonsai-27B-gguf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "prism-ml/Ternary-Bonsai-27B-gguf"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "prism-ml/Ternary-Bonsai-27B-gguf",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/prism-ml/Ternary-Bonsai-27B-gguf:F16

Ollama
How to use prism-ml/Ternary-Bonsai-27B-gguf with Ollama:
```
ollama run hf.co/prism-ml/Ternary-Bonsai-27B-gguf:F16
```
Unsloth Desktop

How to use prism-ml/Ternary-Bonsai-27B-gguf with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf prism-ml/Ternary-Bonsai-27B-gguf:F16

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "prism-ml/Ternary-Bonsai-27B-gguf:F16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use prism-ml/Ternary-Bonsai-27B-gguf with Docker Model Runner:
```
docker model run hf.co/prism-ml/Ternary-Bonsai-27B-gguf:F16
```

Lemonade

How to use prism-ml/Ternary-Bonsai-27B-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull prism-ml/Ternary-Bonsai-27B-gguf:F16

Run and chat with the model

lemonade run user.Ternary-Bonsai-27B-gguf-F16

List all available models

lemonade list

Hermes Agent

How to use prism-ml/Ternary-Bonsai-27B-gguf with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf prism-ml/Ternary-Bonsai-27B-gguf:F16

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default prism-ml/Ternary-Bonsai-27B-gguf:F16

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use prism-ml/Ternary-Bonsai-27B-gguf with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf prism-ml/Ternary-Bonsai-27B-gguf:F16

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "prism-ml/Ternary-Bonsai-27B-gguf:F16" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

New discussion

Resources

View closed (8)

My hands-on notes from 5 hard tasks: math/security/coding are strong; two patterns worth a look (thinking loops, unchecked claims)

#63 opened 3 days ago by

jfriedler

🚀 Starting Agent Loop Tool Efficiency Test

#62 opened 4 days ago by

bukit

Inference demo does not work.

#61 opened 6 days ago by

ZeroWw

Serious Multilingual function errors

#60 opened 12 days ago by

claros-it

Benchmark: 27B Q2_0 vs Qwen3.6-27B IQ2_XXS on one RTX 5060 Ti 16 GB (quality@budget, speed, VRAM, energy)

❤️ 2

#58 opened 25 days ago by

Astezelex

How to Run this model Properly, Tested

🔥 3

#57 opened about 1 month ago by

YahiaAbdelraoof

RTX 3060 12 GB: 31 tok/s on small contexts

#55 opened about 1 month ago by

d9k

Will the qwen3.8-27b version be produced?

🚀➕ 28

#54 opened about 1 month ago by

cleverhack

PLEASE COME BACK😭😭😭😭

🚀❤️ 24

#53 opened about 1 month ago by

AsThirtyThree

Qwen3.8-27b version

❤️ 44

#52 opened about 1 month ago by

Exo87

Speed when CPU-only?

#51 opened about 1 month ago by

TriAxp

Update metadata (tokenizer.ggml.model)

#50 opened about 2 months ago by

Lethaulte

Prefix reuse works with the mmproj loaded - the launcher flags disable it (102s -> 0.14s on a 178k prompt, upstream llama.cpp)

#49 opened about 2 months ago by

EngineerReversed

A comprehensive review

👍 1

#48 opened about 2 months ago by

gdevenyi

Appreciation Post

🤗👀 3

#47 opened 2 months ago by

Maani

Very promising model

#46 opened 2 months ago by

Mk2Oracle

Ternary version of Laguna-S-2.1 118B A8B?

👍 1

#45 opened 2 months ago by

gmm89m

can do to gemma-3-27b too? that model is really great for writing and language

#44 opened 2 months ago by

Hakemz

MOE or changing to A3B

#43 opened 2 months ago by

Tom-Neverwinter

Multi-Token Prediction version 🙏

🔥 2

#42 opened 2 months ago by

Alexzander85

loopy, hallucinates and tool calling problems

#41 opened 2 months ago by

lobstertot

how to partially fix t/s speed generation and partially unlock Dspark full speed capabilities

#40 opened 2 months ago by

LinkuStarto

Ternary-Bonsai-27B-Q2_g64.gguf E gguf_init_from_reader: tensor 'output.weight' has invalid ggml type 42. should be in [0, 42)

#39 opened 2 months ago by

Petrova-1

GLM 5.2

🔥 11

#37 opened 2 months ago by

Kizna

Beware of the kv_cache params!

😔🔥 7

#36 opened 2 months ago by

aaron-newsome

That's an odd observation.

#35 opened 2 months ago by

chunkuk

Breaking occurs in Korean or other East Asian languages.

#34 opened 2 months ago by

chunkuk

error lmstudio

#32 opened 2 months ago by

Metromax

Training Cost

#31 opened 2 months ago by

CurbStomper

Repetition on Inference Providers both on Hugging Face and demo.

#30 opened 2 months ago by

oro872gioioso

Please publish more models

🤗 4

#29 opened 2 months ago by

swiftarrow

Unable to pull Q8_0

#26 opened 2 months ago by

ly2025

vLLM kernels question + RAG concurrency analysis

#25 opened 2 months ago by

Astezelex

Benchmark: 27B Q2_0 vs Qwen3.6-27B IQ2_XXS on one RTX 5060 Ti 16 GB (quality@budget, speed, VRAM, energy)

❤️🔥 7

#24 opened 2 months ago by

Astezelex

Thank you for enabling local models to be useful! - MoE models next please!

🔥 12

#23 opened 2 months ago by

KhaledAshour

Failed to load the model

#22 opened 2 months ago by

parseco

NPU?

#21 opened 2 months ago by

neoOpus

Script to convert PQ2_0 to Q2_0 - code in-line here

#20 opened 2 months ago by

sfrav

Generates gibberish and repeats itself in the middle of thinking and token generation

🚀👀 2

#19 opened 2 months ago by

rbtasd

What parameters can generally be set in LM Studio to achieve optimal performance?

#18 opened 2 months ago by

pymo

turboquant for 2 bit model https://github.com/jarkevithwlad/turboquant-prismml-cuda

❤️👍 7

#17 opened 2 months ago by

jarkevithwlad

10 tok/s on an RTX 3060 12GB (Q2_0)

🔥 4

#16 opened 2 months ago by

bukit

it hallicunates hard

🚀 1

#15 opened 2 months ago by

AsThirtyThree

Ternary-Bonsai-27B Q2_0 works on ROCm / RDNA3 (RX 7800 XT, gfx1101) — benchmarks + iGPU gotcha

🔥👍 3

#14 opened 2 months ago by

basselhajj

We need bonsai gemma 4 12B

👍 2

#13 opened 2 months ago by

CNWPlayer

Unable to get D-Spark Q4_1 working with Q2_0 Model file

#11 opened 2 months ago by

AceTheUnfazed

Documenting what each GGUF file does

👍 10

#10 opened 2 months ago by

croqaz

Cost for quantisation?

👀➕ 5

#9 opened 2 months ago by

Luke2642

approved online demo?

#8 opened 2 months ago by

ct-2

Please share this research to be replicable

🔥 9

#7 opened 2 months ago by

CYISNOTHERE