Instructions to use HaithamalWaisy/kimi-k3-merged with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use HaithamalWaisy/kimi-k3-merged 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 HaithamalWaisy/kimi-k3-merged
# Run inference directly in the terminal:
llama cli -hf HaithamalWaisy/kimi-k3-merged

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf HaithamalWaisy/kimi-k3-merged
# Run inference directly in the terminal:
llama cli -hf HaithamalWaisy/kimi-k3-merged

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 HaithamalWaisy/kimi-k3-merged
# Run inference directly in the terminal:
./llama-cli -hf HaithamalWaisy/kimi-k3-merged

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 HaithamalWaisy/kimi-k3-merged
# Run inference directly in the terminal:
./build/bin/llama-cli -hf HaithamalWaisy/kimi-k3-merged

Use Docker

docker model run hf.co/HaithamalWaisy/kimi-k3-merged

LM Studio
Jan
Ollama
How to use HaithamalWaisy/kimi-k3-merged with Ollama:
```
ollama run hf.co/HaithamalWaisy/kimi-k3-merged
```
Unsloth Desktop

How to use HaithamalWaisy/kimi-k3-merged with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf HaithamalWaisy/kimi-k3-merged

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": "HaithamalWaisy/kimi-k3-merged"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use HaithamalWaisy/kimi-k3-merged with Docker Model Runner:
```
docker model run hf.co/HaithamalWaisy/kimi-k3-merged
```

Lemonade

How to use HaithamalWaisy/kimi-k3-merged with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull HaithamalWaisy/kimi-k3-merged

Run and chat with the model

lemonade run user.kimi-k3-merged-{{QUANT_TAG}}

List all available models

lemonade list

Hermes Agent

How to use HaithamalWaisy/kimi-k3-merged with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf HaithamalWaisy/kimi-k3-merged

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 HaithamalWaisy/kimi-k3-merged

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use HaithamalWaisy/kimi-k3-merged with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf HaithamalWaisy/kimi-k3-merged

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 "HaithamalWaisy/kimi-k3-merged" \
  --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"

Kimi K3 — Merged 2-Expert Model (66.3 GB)

Fully dense, 62.23 B-parameter model produced by collapsing all 896 routed experts of each MoE layer of Kimi K3 into 2 super-experts using the MergeMoE expert-output-merging technique.

Derived from: Moonshot AI's Kimi K3 (open weights, Apache 2.0) via Unsloth's 1-bit GGUF conversion (UD-IQ1_S, 14 shards).

Paper: Kimi K3 Under Compute Constraints: Implementing MergeMoE to Take 14 Shards to 5GB on Consumer Hardware

Author: Haitham al-Waisy · Contact: haithamalwaisy@gmail.com · X/Twitter · YouTube

Code: https://github.com/Haitham-alwaisy/Kimi-K3-MergeMoE

Model anatomy

93 blocks, all 896 experts per layer → 2 super-experts (usage-frequency-weighted average)
All-Q8_0/F32 weights (1,392 Q8_0 + 1,181 F32 tensors)
62.23 B parameters, 100% active per token

Intended use

This model is intended as a base for fine-tuning and Reinforcement Learning. It is recommended to be run inside the Prime Agent harness.

Usage

hf download HaithamalWaisy/kimi-k3-merged kimi-k3-merged-2exp.gguf --local-dir .

License

Apache 2.0 (inherited from Kimi K3). Quality was not benchmarked by this project.

Downloads last month: 844

GGUF

Model size

62B params

Architecture

kimi-k3

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support