Instructions to use baa-ai/Kimi-K2.6-RAM-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use baa-ai/Kimi-K2.6-RAM-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="baa-ai/Kimi-K2.6-RAM-GGUF",
	filename="kimi-k2.6-ram-344gb.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use baa-ai/Kimi-K2.6-RAM-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf baa-ai/Kimi-K2.6-RAM-GGUF
# Run inference directly in the terminal:
llama-cli -hf baa-ai/Kimi-K2.6-RAM-GGUF

Install from WinGet (Windows)

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

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 baa-ai/Kimi-K2.6-RAM-GGUF
# Run inference directly in the terminal:
./llama-cli -hf baa-ai/Kimi-K2.6-RAM-GGUF

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

Use Docker

docker model run hf.co/baa-ai/Kimi-K2.6-RAM-GGUF

LM Studio
Jan

vLLM

How to use baa-ai/Kimi-K2.6-RAM-GGUF with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/baa-ai/Kimi-K2.6-RAM-GGUF

Ollama
How to use baa-ai/Kimi-K2.6-RAM-GGUF with Ollama:
```
ollama run hf.co/baa-ai/Kimi-K2.6-RAM-GGUF
```

Unsloth Studio new

How to use baa-ai/Kimi-K2.6-RAM-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for baa-ai/Kimi-K2.6-RAM-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for baa-ai/Kimi-K2.6-RAM-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for baa-ai/Kimi-K2.6-RAM-GGUF to start chatting

Pi new

How to use baa-ai/Kimi-K2.6-RAM-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf baa-ai/Kimi-K2.6-RAM-GGUF

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "baa-ai/Kimi-K2.6-RAM-GGUF"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use baa-ai/Kimi-K2.6-RAM-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf baa-ai/Kimi-K2.6-RAM-GGUF

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 baa-ai/Kimi-K2.6-RAM-GGUF

Run Hermes

hermes

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

Lemonade

How to use baa-ai/Kimi-K2.6-RAM-GGUF with Lemonade:

Pull the model

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

Run and chat with the model

lemonade run user.Kimi-K2.6-RAM-GGUF-{{QUANT_TAG}}

List all available models

lemonade list

Kimi-K2.6 RAM GGUF

Mixed-precision GGUF quantizations of moonshotai/Kimi-K2.6, produced with the RAM (Resource-Aware Mixed-precision) pipeline.

Each variant is also published as its own repo for easier pinning:

baa-ai/Kimi-K2.6-RAM-344GB-GGUF — Q2_K experts, lower-footprint
baa-ai/Kimi-K2.6-RAM-447GB-GGUF — Q3_K experts, higher-quality

Files

File	Size	Expert bits	Attention bits	Target hardware
`kimi-k2.6-ram-344gb.gguf`	344 GB	Q2_K	Q5_K–Q8_0 (probe-allocated)	2× 192 GB or 1× 512 GB
`kimi-k2.6-ram-447gb.gguf`	447 GB	Q3_K	Q5_K–Q8_0 (probe-allocated)	2× 256 GB or 1× 512 GB

Method

Quantization bit depths are assigned per-tensor using sensitivity probing rather than a uniform scheme. Each attention tensor receives bits proportional to how much its output diverges under quantization noise, measured across 8 random probes. Expert tensors (384 routed experts × 60 MoE layers) are quantized uniformly at Q2_K or Q3_K depending on the target size.

Architecture: DeepSeek-V3 decoder (kimi_k2), 61 layers, 384 routed experts / 8 active per token, MLA attention, hidden_size=7168, 1.03T total parameters.

Usage (llama.cpp)

llama-cli \
  -m kimi-k2.6-ram-447gb.gguf \
  -c 8192 \
  --temp 0.6 \
  -p "You are a helpful assistant."

Notes

No importance matrix (imatrix) was used — the 1.1 TB Q8_0 intermediate required for imatrix generation exceeds available RAM for inference on this machine. Bit allocation from sensitivity probing provides the primary quality signal.
The original model uses Neural Magic compressed-tensors INT4 for expert weights. These are dequantized and re-quantized to GGUF format during conversion.

Downloads last month: 991

GGUF

Model size

1T params

Architecture

deepseek2

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Model tree for baa-ai/Kimi-K2.6-RAM-GGUF

Base model

moonshotai/Kimi-K2.6

Quantized

(32)

this model