Instructions to use ubergarm/Kimi-K2.6-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 ubergarm/Kimi-K2.6-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 ubergarm/Kimi-K2.6-GGUF:Q2_K
# Run inference directly in the terminal:
llama cli -hf ubergarm/Kimi-K2.6-GGUF:Q2_K

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf ubergarm/Kimi-K2.6-GGUF:Q2_K
# Run inference directly in the terminal:
llama cli -hf ubergarm/Kimi-K2.6-GGUF:Q2_K

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 ubergarm/Kimi-K2.6-GGUF:Q2_K
# Run inference directly in the terminal:
./llama-cli -hf ubergarm/Kimi-K2.6-GGUF:Q2_K

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 ubergarm/Kimi-K2.6-GGUF:Q2_K
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ubergarm/Kimi-K2.6-GGUF:Q2_K

Use Docker

docker model run hf.co/ubergarm/Kimi-K2.6-GGUF:Q2_K

LM Studio
Jan

vLLM

How to use ubergarm/Kimi-K2.6-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ubergarm/Kimi-K2.6-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": "ubergarm/Kimi-K2.6-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ubergarm/Kimi-K2.6-GGUF:Q2_K

Ollama
How to use ubergarm/Kimi-K2.6-GGUF with Ollama:
```
ollama run hf.co/ubergarm/Kimi-K2.6-GGUF:Q2_K
```
Unsloth Desktop

How to use ubergarm/Kimi-K2.6-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ubergarm/Kimi-K2.6-GGUF:Q2_K

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": "ubergarm/Kimi-K2.6-GGUF:Q2_K"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use ubergarm/Kimi-K2.6-GGUF with Docker Model Runner:
```
docker model run hf.co/ubergarm/Kimi-K2.6-GGUF:Q2_K
```

Lemonade

How to use ubergarm/Kimi-K2.6-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ubergarm/Kimi-K2.6-GGUF:Q2_K

Run and chat with the model

lemonade run user.Kimi-K2.6-GGUF-Q2_K

List all available models

lemonade list

Hermes Agent

How to use ubergarm/Kimi-K2.6-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 ubergarm/Kimi-K2.6-GGUF:Q2_K

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 ubergarm/Kimi-K2.6-GGUF:Q2_K

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use ubergarm/Kimi-K2.6-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ubergarm/Kimi-K2.6-GGUF:Q2_K

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 "ubergarm/Kimi-K2.6-GGUF:Q2_K" \
  --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 2.7 GGUF?

#14

by ChuckMcSneed - opened Jun 14

Discussion

ChuckMcSneed

Jun 14

Are you planning to make ggufs for new kimi https://huggingface.co/moonshotai/Kimi-K2.7-Code ?

Thireus

Jun 16

•

edited Jun 16

@ubergarm , I need to know too :D eagerly waiting for an imatrix. I've converted the model to BF16 here: https://huggingface.co/Thireus/Kimi-K2.7-Code-THIREUS-BF16-SPECIAL_SPLIT

Fernanda24

Jun 17

and GLM 5.2 🖐

Thireus

Jun 17

BF16 for GLM-5.2 is here: https://huggingface.co/Thireus/GLM-5.2-THIREUS-BF16-SPECIAL_SPLIT

ubergarm

Owner Jun 22

Sorry guys, I still don't have access the the big remote rig after some network maintenance so taking some time off for the summer. Kimi-K2.7 and GLM-5.2 are looking hot though so sad I'm not releasing at the moment! 😥

I'll keep y'all posted as I get more updates myself. Cheers and be well!

Fernanda24

Jun 24

we need the best quality ggufs back, what kind of rig specs do you need and how much time generally for a model like glm 5.2 or kimi k2.7 to produce and upload all the quant variations?

ubergarm

Owner Jul 10

@Fernanda24

I've been using a big dual socket AMD EPYC rig with almost 1.5TB RAM and zero GPU/s or VRAM. Running everything on CPU is fine it just takes some time. The main resource requirement is enough RAM to fit the entire bf16 (or q8_0 depending on how the original safetensors was released) to make the importance matrix. Once you have that, it doesn't take much but a lot of disk space and network to make the quants and upload to hf. The other issue is most folks run out of huggingface storage very quickly with these big quants.

Fortunately, some other people are picking up my slack and @muzzy released a GLM-5.2 https://huggingface.co/muzzy/GLM-5.2-GGUF and check for the ik_llama.cpp tag on huggingface. I've not tried the latest GLM MTP stuff on ik.

gghfez

Jul 13

The main resource requirement is enough RAM to fit the entire bf16 (or q8_0 depending on how the original safetensors was released) to make the importance matrix.

Fortunately, some other people are picking up my slack and @muzzy

I made a script a while back to pinch mainline llama.cpp imatrix.gguf files and convert them to the ik_llama.cpp imatrix.dat format.

I've converted it into a HF space: gghfez/ik_llama_imatrix_converter

I've done a few quants locally using Unsloth, Bartowski and AesSedai imatrix files and it's worked well so far.

Only caveat is there might be issues with models like Mimo-V2.5-Pro where mainline fuse the atten q, k, v

I intend to use and maintain it if anything breaks when new models are released.

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