Instructions to use picklass/Kimi-K2-Thinking-MLX-4.25bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use picklass/Kimi-K2-Thinking-MLX-4.25bit with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("picklass/Kimi-K2-Thinking-MLX-4.25bit")

prompt = "Write a story about Einstein"
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

text = generate(model, tokenizer, prompt=prompt, verbose=True)

Notebooks
Google Colab
Kaggle
Local Apps Settings
LM Studio

How to use picklass/Kimi-K2-Thinking-MLX-4.25bit with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "picklass/Kimi-K2-Thinking-MLX-4.25bit"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "picklass/Kimi-K2-Thinking-MLX-4.25bit"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use picklass/Kimi-K2-Thinking-MLX-4.25bit with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "picklass/Kimi-K2-Thinking-MLX-4.25bit"

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 picklass/Kimi-K2-Thinking-MLX-4.25bit

Run Hermes

hermes

MLX LM

How to use picklass/Kimi-K2-Thinking-MLX-4.25bit with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "picklass/Kimi-K2-Thinking-MLX-4.25bit"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "picklass/Kimi-K2-Thinking-MLX-4.25bit"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "picklass/Kimi-K2-Thinking-MLX-4.25bit",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

NOTICE

This model has been superseded by the higher quality Kimi-K2.6-Q3.5-INF version available here

INFORMATION

See Kimi-K2-Thinking 4.25bit MLX in action - demonstration video

q4.25bit quant perplexity TBA, but q4.5bit quant typically achieves 1.168 perplexity in our testing

Quantization	Perplexity
q2.5	41.293
q3.5	1.900
q3.95	1.243
q4.25	TBA
q4.5	1.168
q6.5	1.128
q8.5	1.128

Usage Notes

Tested on a M3 Ultra 512GB RAM connected to MBP 128GB RAM using Inferencer app v1.6 with distributed compute
For more information on the distributed compute feature see: github.com/inferencer/issues/31
Memory usage: MBP ~80GB + Mac Studio ~450GB
Expect ~22 tokens/s @ 1000 tokens
Quantized with a modified version of MLX 0.28
For more details see demonstration video or visit Kimi-K2-Thinking.

Disclaimer

We are not the creator, originator, or owner of any model listed. Each model is created and provided by third parties. Models may not always be accurate or contextually appropriate. You are responsible for verifying the information before making important decisions. We are not liable for any damages, losses, or issues arising from its use, including data loss or inaccuracies in AI-generated content.

Downloads last month: 302

Model tree for picklass/Kimi-K2-Thinking-MLX-4.25bit

Base model

moonshotai/Kimi-K2-Thinking

Quantized

(19)

this model