Instructions to use abenzerps/K2-Horizon-7B-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 abenzerps/K2-Horizon-7B-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 abenzerps/K2-Horizon-7B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf abenzerps/K2-Horizon-7B-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 abenzerps/K2-Horizon-7B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf abenzerps/K2-Horizon-7B-GGUF:Q4_K_M

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 abenzerps/K2-Horizon-7B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf abenzerps/K2-Horizon-7B-GGUF:Q4_K_M

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 abenzerps/K2-Horizon-7B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf abenzerps/K2-Horizon-7B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/abenzerps/K2-Horizon-7B-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use abenzerps/K2-Horizon-7B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "abenzerps/K2-Horizon-7B-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": "abenzerps/K2-Horizon-7B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/abenzerps/K2-Horizon-7B-GGUF:Q4_K_M

Ollama
How to use abenzerps/K2-Horizon-7B-GGUF with Ollama:
```
ollama run hf.co/abenzerps/K2-Horizon-7B-GGUF:Q4_K_M
```
Unsloth Desktop

How to use abenzerps/K2-Horizon-7B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf abenzerps/K2-Horizon-7B-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": "abenzerps/K2-Horizon-7B-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use abenzerps/K2-Horizon-7B-GGUF with Docker Model Runner:
```
docker model run hf.co/abenzerps/K2-Horizon-7B-GGUF:Q4_K_M
```

Lemonade

How to use abenzerps/K2-Horizon-7B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull abenzerps/K2-Horizon-7B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.K2-Horizon-7B-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use abenzerps/K2-Horizon-7B-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 abenzerps/K2-Horizon-7B-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 abenzerps/K2-Horizon-7B-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use abenzerps/K2-Horizon-7B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf abenzerps/K2-Horizon-7B-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 "abenzerps/K2-Horizon-7B-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"

K2-Horizon-7B-GGUF

File size: 2,849 Bytes

---
base_model: IFM/K2-Horizon-7B
base_model_relation: quantized
license: apache-2.0
language:
  - en
library_name: gguf
pipeline_tag: text-generation
tags:
  - gguf
  - llama.cpp
  - k2-horizon
  - long-context
  - 512k-context
  - dense
---

> [!IMPORTANT]
> **Compatibility:** These GGUF files require a `llama.cpp` build with K2 Horizon architecture support. Until upstream support lands, use the [MBZUAI-IFM fork](https://github.com/MBZUAI-IFM/llama.cpp/tree/model/K2Horizon).

# K2-Horizon-7B GGUF

GGUF quantizations of [IFM/K2-Horizon-7B](https://huggingface.co/IFM/K2-Horizon-7B), a 7B dense decoder-only model for reasoning, coding, long-context work, and tool use. The source checkpoint supports a native context length of **524,288 tokens (512K)**.

## Benchmarks

![K2-Horizon-7B benchmark results](assets/k2-horizon-7b-benchmarks.png)

*Benchmark results reported by IFM for the original K2-Horizon-7B checkpoint.*

## GGUF files

| Quantization | File | Size |
| --- | --- | ---: |
| Q4_0 | [K2-Horizon-7B-Q4_0.gguf](K2-Horizon-7B-Q4_0.gguf) | 5.34 GB |
| Q4_K_M | [K2-Horizon-7B-Q4_K_M.gguf](K2-Horizon-7B-Q4_K_M.gguf) | 5.59 GB |
| Q4_K_M Selective | [K2-Horizon-7B-Q4_K_M-Selective.gguf](K2-Horizon-7B-Q4_K_M-Selective.gguf) | 5.96 GB |
| Q5_K_M | [K2-Horizon-7B-Q5_K_M.gguf](K2-Horizon-7B-Q5_K_M.gguf) | 6.47 GB |
| Q6_K | [K2-Horizon-7B-Q6_K.gguf](K2-Horizon-7B-Q6_K.gguf) | 7.39 GB |
| Q8_0 | [K2-Horizon-7B-Q8_0.gguf](K2-Horizon-7B-Q8_0.gguf) | 9.57 GB |

The files are text-only GGUFs; no vision projector is required. The selective variant uses a Q4_K_M baseline with attention Q/K/V/O projection tensors kept at Q6_K; it is a manual tensor-selective build and does not use an importance matrix. SHA-256 checksums are provided in [`SHA256SUMS.txt`](SHA256SUMS.txt).

## Chat template

Each GGUF embeds the llama.cpp-compatible chat template. [`chat_template.jinja`](chat_template.jinja) is a matching external copy for tools that require one. The original source template is retained as [`chat_template.upstream.jinja`](chat_template.upstream.jinja) for runtimes with full Jinja support.

`xml` is the default tool-call format. Use `--chat-template-kwargs` to select `json` or `xml_typed` when required.

## Usage

Use the [IFM K2 Horizon llama.cpp fork](https://github.com/MBZUAI-IFM/llama.cpp/tree/model/K2Horizon). The example below uses a practical 128K context; `-c 524288` can be used when the available memory is sufficient.

```bash
llama-cli \
  -m K2-Horizon-7B-Q4_K_M.gguf \
  -c 131072 --jinja \
  --temp 1.0 --top-p 0.95
```

## Source

- Source model: [IFM/K2-Horizon-7B](https://huggingface.co/IFM/K2-Horizon-7B)
- Source revision: [`2c9659a`](https://huggingface.co/IFM/K2-Horizon-7B/commit/2c9659a84c4eea6f9f60462221fe762c8c84d75c)
- Source license: [Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0)