Instructions to use unsloth/Kimi-K2-Thinking-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use unsloth/Kimi-K2-Thinking-GGUF with Transformers:

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("unsloth/Kimi-K2-Thinking-GGUF", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use unsloth/Kimi-K2-Thinking-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 unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL
# Run inference directly in the terminal:
llama cli -hf unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL
# Run inference directly in the terminal:
llama cli -hf unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL

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 unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL
# Run inference directly in the terminal:
./llama-cli -hf unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL

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 unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL
# Run inference directly in the terminal:
./build/bin/llama-cli -hf unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL

Use Docker

docker model run hf.co/unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL

LM Studio
Jan
Ollama
How to use unsloth/Kimi-K2-Thinking-GGUF with Ollama:
```
ollama run hf.co/unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL
```
Unsloth Desktop

How to use unsloth/Kimi-K2-Thinking-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL

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": "unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use unsloth/Kimi-K2-Thinking-GGUF with Docker Model Runner:
```
docker model run hf.co/unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL
```

Lemonade

How to use unsloth/Kimi-K2-Thinking-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL

Run and chat with the model

lemonade run user.Kimi-K2-Thinking-GGUF-UD-Q4_K_XL

List all available models

lemonade list

Hermes Agent

How to use unsloth/Kimi-K2-Thinking-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 unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL

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 unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use unsloth/Kimi-K2-Thinking-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL

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 "unsloth/Kimi-K2-Thinking-GGUF:UD-Q4_K_XL" \
  --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"

Chat template fixes + FAQ

pinned

by danielhanchen - opened Nov 8, 2025

Discussion

danielhanchen

Unsloth AI org Nov 8, 2025

•

edited Nov 8, 2025

Hey everyone! We worked with the Kimi K2 team to fix the default system prompt of You are Kimi, an AI assistant created by Moonshot AI. not appearing in the first turn! See https://huggingface.co/moonshotai/Kimi-K2-Thinking/discussions/12 for more details.

All GGUFs, imatrix calibrated GGUFs and uploads are all updated with these fixes.

You may notice that there are no thinking tags when you run the model. This is normal and intended behavior. To enable it you will need to add --special.

Don't forget to read our guide for more details: https://docs.unsloth.ai/models/kimi-k2-thinking-how-to-run-locally

danielhanchen pinned discussion Nov 8, 2025

danielhanchen changed discussion title from Chat template and bug fixes to Chat template bug fixes Nov 8, 2025

danielhanchen changed discussion title from Chat template bug fixes to Chat template fixes + Q&A Nov 8, 2025

mtcl

Nov 9, 2025

Would you know why i get this "< | im_end | >" at the end of the message? Apart from that everything works perfectly.

Today at 5:05 PM
Thought for 14 seconds
Hi! How can I help you today?<|im_end|>

adsfdgfhgjhk11

Nov 10, 2025

Have you been able to evaluate drop off in accuracy for the Q2 UD ? Or whichever?

danielhanchen changed discussion title from Chat template fixes + Q&A to Chat template fixes + FAQ Nov 12, 2025

danielhanchen

Unsloth AI org Nov 12, 2025

Would you know why i get this "< | im_end | >" at the end of the message? Apart from that everything works perfectly.

Today at 5:05 PM
Thought for 14 seconds
Hi! How can I help you today?<|im_end|>

Hello we wrote more info in our guide: https://docs.unsloth.ai/models/kimi-k2-thinking-how-to-run-locally#no-thinking-tags

CHNtentes

Nov 12, 2025

•

edited Nov 12, 2025

Would you know why i get this "< | im_end | >" at the end of the message? Apart from that everything works perfectly.

Today at 5:05 PM
Thought for 14 seconds
Hi! How can I help you today?<|im_end|>

Hello we wrote more info in our guide: https://docs.unsloth.ai/models/kimi-k2-thinking-how-to-run-locally#no-thinking-tags

You could create a page for some basic knowledge about LLM in general, like how tokenizer and chat template works.

rayw2k25

Nov 12, 2025

You can remove the artifacts using the chat template.

mtcl

Nov 12, 2025

i was able to fix this, here is the complete command for anyone else looking for the solution:

CUDA_DEVICE_ORDER=PCI_BUS_ID CUDA_VISIBLE_DEVICES="0,1" ./build/bin/llama-server \
  --model /media/data/models/unsloth/Kimi-K2-Thinking-GGUF/UD-Q2_K_XL/Kimi-K2-Thinking-UD-Q2_K_XL-00001-of-00008.gguf \
  --alias unsloth/Kimi-K2-Thinking \
  --ctx-size 262144 \
  -fa on \
  -b 4096 -ub 4096 \
  -ngl 99 \
  -ot exps=CPU \
  --special \
  --no-warmup \
  --parallel 1 \
  --temp 1.0 \
  --min-p 0.01 \
  --jinja \
  --threads 56 \
  --threads-batch 56 \
  --host 0.0.0.0 \
  --port 10002 \
  --reverse-prompt "<|im_end|>"

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment