Instructions to use bartowski/K2-Horizon-MoVA-36B-A4B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use bartowski/K2-Horizon-MoVA-36B-A4B-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 bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M

Llama new
LM Studio
Jan

vLLM

How to use bartowski/K2-Horizon-MoVA-36B-A4B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "bartowski/K2-Horizon-MoVA-36B-A4B-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": "bartowski/K2-Horizon-MoVA-36B-A4B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M

Ollama
How to use bartowski/K2-Horizon-MoVA-36B-A4B-GGUF with Ollama:
```
ollama run hf.co/bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M
```
Unsloth Desktop

How to use bartowski/K2-Horizon-MoVA-36B-A4B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M

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": "bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use bartowski/K2-Horizon-MoVA-36B-A4B-GGUF with Docker Model Runner:
```
docker model run hf.co/bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M
```

Lemonade

How to use bartowski/K2-Horizon-MoVA-36B-A4B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.K2-Horizon-MoVA-36B-A4B-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use bartowski/K2-Horizon-MoVA-36B-A4B-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 bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M

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 bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use bartowski/K2-Horizon-MoVA-36B-A4B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M

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 "bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M" \
  --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"

Llamacpp imatrix Quantizations of K2-Horizon-MoVA-36B-A4B by IFM

Using llama.cpp release b11462 for quantization.

Original model: https://huggingface.co/IFM/K2-Horizon-MoVA-36B-A4B

Model details:

Parameter count: 37B (source checkpoint)
Input support: text
Speculative decoding: no
imatrix: yes - details

How to run

Prompt format

<|ifm|begin_of_text|><|ifm|im_start|>system
{system_prompt}<|ifm|im_end|><|ifm|im_start|>user
{prompt}<|ifm|im_end|><|ifm|im_start|>assistant
<ifm|think>

Prompt format with tool definitions

<|ifm|begin_of_text|><|ifm|im_start|>system
# Tools
You may call one or more tools to assist with the user query.

Available tools are:

<ifm|tools>
## get_stock_price
Get the current stock price

**Parameters**
- `symbol` *(string, required)* - The stock symbol, e.g. AAPL, GOOG
</ifm|tools>

When calling tools, you MUST follow the tool-call format below:

Wrap all tool calls in a single <ifm|tool_calls></ifm|tool_calls> block. For each call, write the function name at the start of <ifm|tool_call>, followed by paired <ifm|arg_key> and <ifm|arg_value> tags for each argument:

<ifm|tool_calls>
<ifm|tool_call>$FUNCTION_NAME
<ifm|arg_key>$PARAMETER_NAME</ifm|arg_key>
<ifm|arg_value>$PARAMETER_VALUE</ifm|arg_value>
...
</ifm|tool_call>
</ifm|tool_calls>

String and scalar parameters should be written as plain text. Array and object parameters should be written as JSON literals.

{system_prompt}<|ifm|im_end|><|ifm|im_start|>user
{prompt}<|ifm|im_end|><|ifm|im_start|>assistant
<ifm|think>

Don't know which to choose? Grab Q4_K_M (22.70GB) - usually a good mix of size and performance. Download instructions available here

Available files:

Filename	Quant type	File Size	Split	Description
K2-Horizon-MoVA-36B-A4B-bf16.gguf	bf16	74.92GB	true	Full BF16 weights.
K2-Horizon-MoVA-36B-A4B-Q8_0.gguf	Q8_0	39.83GB	false	Extremely high quality, generally unneeded but max available quant.
K2-Horizon-MoVA-36B-A4B-Q6_K_L.gguf	Q6_K_L	32.40GB	false	Uses Q8_0 for embed and output weights. Very high quality, near perfect, recommended.
K2-Horizon-MoVA-36B-A4B-Q6_K.gguf	Q6_K	32.09GB	false	Very high quality, near perfect, recommended.
K2-Horizon-MoVA-36B-A4B-Q5_K_M.gguf	Q5_K_M	26.71GB	false	High quality, recommended.
K2-Horizon-MoVA-36B-A4B-Q5_K_S.gguf	Q5_K_S	26.03GB	false	High quality, recommended.
K2-Horizon-MoVA-36B-A4B-Q4_1.gguf	Q4_1	23.76GB	false	Legacy format, similar performance to Q4_K_S but with improved tokens/watt on Apple silicon.
K2-Horizon-MoVA-36B-A4B-Q4_K_L.gguf	Q4_K_L	23.18GB	false	Uses Q8_0 for embed and output weights. Good quality, recommended.
K2-Horizon-MoVA-36B-A4B-Q4_K_M.gguf	Q4_K_M	22.70GB	false	Good quality, default size for most use cases, recommended.
K2-Horizon-MoVA-36B-A4B-Q4_K_S.gguf	Q4_K_S	22.04GB	false	Slightly lower quality with more space savings, recommended.
K2-Horizon-MoVA-36B-A4B-Q4_0.gguf	Q4_0	21.56GB	false	Legacy format, kept for compatibility with older tools.
K2-Horizon-MoVA-36B-A4B-IQ4_NL.gguf	IQ4_NL	21.54GB	false	Similar to IQ4_XS, but slightly larger.
K2-Horizon-MoVA-36B-A4B-IQ4_XS.gguf	IQ4_XS	20.42GB	false	Decent quality, smaller than Q4_K_S with similar performance, recommended.
K2-Horizon-MoVA-36B-A4B-IQ3_M.gguf	IQ3_M	17.93GB	false	Medium-low quality, new method with decent performance comparable to Q3_K_M.
K2-Horizon-MoVA-36B-A4B-Q3_K_L.gguf	Q3_K_L	17.82GB	false	Lower quality but usable, good for low RAM availability.
K2-Horizon-MoVA-36B-A4B-Q3_K_M.gguf	Q3_K_M	17.35GB	false	Low quality.
K2-Horizon-MoVA-36B-A4B-IQ3_XS.gguf	IQ3_XS	17.32GB	false	Lower quality, new method with decent performance, slightly better than Q3_K_S.
K2-Horizon-MoVA-36B-A4B-Q3_K_S.gguf	Q3_K_S	16.76GB	false	Low quality, not recommended.
K2-Horizon-MoVA-36B-A4B-IQ3_XXS.gguf	IQ3_XXS	15.80GB	false	Lower quality, new method with decent performance, comparable to Q3 quants.
K2-Horizon-MoVA-36B-A4B-Q2_K.gguf	Q2_K	13.57GB	false	Very low quality but surprisingly usable.
K2-Horizon-MoVA-36B-A4B-IQ2_M.gguf	IQ2_M	13.09GB	false	Relatively low quality, uses SOTA techniques to be surprisingly usable.
K2-Horizon-MoVA-36B-A4B-IQ2_S.gguf	IQ2_S	11.97GB	false	Low quality, uses SOTA techniques to be usable.
K2-Horizon-MoVA-36B-A4B-IQ2_XS.gguf	IQ2_XS	11.72GB	false	Low quality, uses SOTA techniques to be usable.
K2-Horizon-MoVA-36B-A4B-IQ2_XXS.gguf	IQ2_XXS	10.64GB	false	Very low quality, uses SOTA techniques to be usable.

Download a specific file:

hf download bartowski/K2-Horizon-MoVA-36B-A4B-GGUF --include "K2-Horizon-MoVA-36B-A4B-Q4_K_M.gguf" --local-dir ./

Downloading using the Hugging Face CLI

Click to view download instructions

First, make sure you have the Hugging Face CLI installed:

pip install -U "huggingface_hub[cli]"

Download a specific file:

hf download bartowski/K2-Horizon-MoVA-36B-A4B-GGUF --include "K2-Horizon-MoVA-36B-A4B-Q4_K_M.gguf" --local-dir ./

The files marked true in the Split column above are stored as multiple parts in a folder. To download all the parts to a local folder, run:

hf download bartowski/K2-Horizon-MoVA-36B-A4B-GGUF --include "K2-Horizon-MoVA-36B-A4B-bf16/*" --local-dir ./

You can either specify a new local-dir (K2-Horizon-MoVA-36B-A4B-bf16) or download them all in place (./)

How to run

These quants run with llama.cpp - installable in one line via llama.app:

curl -LsSf https://llama.app/install.sh | sh
llama-server -hf bartowski/K2-Horizon-MoVA-36B-A4B-GGUF:Q4_K_M

llama-server includes a built-in chat web UI, served at http://localhost:8080 by default.

These quants were made with llama.cpp release b11462 - if this model's architecture is newly supported, you'll need that release or newer to run them.

They also work in: LM Studio · koboldcpp · ramalama · Jan AI · Text Generation Web UI · LoLLMs · Atomic Chat

imatrix

All quants made using imatrix option, with a calibration corpus rendered through this model's own chat template. The corpus pairs plain prose with tool-calling and reasoning conversations (corpus source data), encoded exactly as this model sees them at inference and processed with --parse-special, so chat-format special tokens contribute to the importance matrix. The corpus rendered for this model is included in this repo: K2-Horizon-MoVA-36B-A4B-calibration-v6.txt. The imatrix is available here: K2-Horizon-MoVA-36B-A4B-imatrix.gguf.

Calibration render details

{
  "recipe": "calibration-v6",
  "encoder": "chat_template",
  "library_versions": {
    "transformers": "5.9.0",
    "tokenizers": "0.22.2",
    "tiktoken": "0.14.0",
    "blobfile": "3.3.0",
    "huggingface_hub": "1.31.0"
  },
  "chunk_size": 512,
  "prose_chunks": 207,
  "tool_chunks": 290,
  "total_chunks": 497,
  "n_conversations": 137,
  "conversation_token_lengths": [
    428,
    1343,
    998,
    1117,
    873,
    1181,
    2744,
    596,
    965,
    1073,
    761,
    1748,
    734,
    992,
    2371,
    998,
    905,
    795,
    615,
    563,
    1183,
    804,
    1042,
    1000,
    1516,
    1246,
    1403,
    731,
    1222,
    1240,
    1208,
    944,
    1026,
    837,
    974,
    1290,
    1371,
    972,
    366,
    1635,
    1136,
    875,
    1051,
    1638,
    1612,
    1084,
    1325,
    642,
    2509,
    924,
    2571,
    605,
    788,
    784,
    837,
    557,
    2096,
    679,
    910,
    774,
    987,
    972,
    713,
    977,
    902,
    1165,
    735,
    1168,
    1762,
    728,
    316,
    834,
    2964,
    2387,
    563,
    717,
    819,
    873,
    1019,
    939,
    896,
    623,
    961,
    829,
    1037,
    1274,
    1087,
    1833,
    672,
    496,
    2337,
    618,
    1074,
    1337,
    1607,
    1044,
    472,
    1156,
    1050,
    1239,
    1511,
    1320,
    626,
    848,
    850,
    2282,
    608,
    526,
    617,
    1104,
    940,
    1235,
    591,
    346,
    308,
    2208,
    839,
    871,
    1423,
    1675,
    2273,
    2263,
    606,
    764,
    674,
    741,
    963,
    792,
    702,
    1067,
    634,
    550,
    1430,
    739,
    704,
    1071,
    1198
  ]
}

Embed/output weights

Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.

ARM/AVX information

llama.cpp automatically "repacks" weights into an interleaved layout at load time for faster inference on ARM and AVX machines - details in this PR. This once required downloading special Q4_0_4_4/4_8/8_8 files; those are long gone. Online repacking now covers Q4_0, IQ4_NL, and most K-quants, so no special quant choice is needed for CPU inference.

Which file should I choose?

Click here for details

An older (early 2024) but still useful write-up with charts comparing quant performances is provided by Artefact2 here

The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have.

If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM.

If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total.

Hugging Face can also do this math for you: add your hardware in your Local Apps settings and the model page will show which files fit.

Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'.

If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M.

If you want to get more into the weeds, you can check out this extremely useful feature chart:

llama.cpp feature matrix

But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size.

These I-quants can also be used on CPU, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide.