Instructions to use magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF",
	filename="Model-LM-IQ2_S.gguf",
)

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

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-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 magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M

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 magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M

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 magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M

Use Docker

docker model run hf.co/magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-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": "magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M

Ollama
How to use magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF with Ollama:
```
ollama run hf.co/magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M
```

Unsloth Studio

How to use magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-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 magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-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 magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF to start chatting

How to use magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M

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": "magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-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 magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M

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 magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M

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 "magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M" \
  --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"

Docker Model Runner
How to use magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF with Docker Model Runner:
```
docker model run hf.co/magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M
```

Lemonade

How to use magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull magiccodingman/Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF-Q4_K_M

List all available models

lemonade list

Qwen3-4B-Instruct-2507-Unsloth-MagicQuant-v2-GGUF / magicquant.replacements.json

magiccodingman

Super-squash branch 'main' using huggingface_hub

23a9647 about 1 month ago

Raw

History Blame Contribute Delete

15.8 kB

	[
	{
	"reasonCode": "SPACING_COLLAPSE",
	"reasonDescription": "Two candidates were too close in practical output space; the stronger one was kept.",
	"rawReason": "meaningful spacing collapse; size gap below 91,013,530 bytes",
	"removed": {
	"key": "0:1:0:3:1:2:102:14:0:0",
	"fileName": "Model-MQ-Q5_K.gguf",
	"displayName": "Model-MQ-Q5_K",
	"shortName": "MQ-Q5_K",
	"provider": "MagicQuant",
	"quantFamily": "Q8_0",
	"isHybrid": true,
	"isExternalPureBaseline": false,
	"kld": 0.006925,
	"ppl": 8.901754,
	"pplDeltaPercent": 0.204354090683967,
	"sizeBytes": 3040771776,
	"sizeGiB": 2.831939399242401
	},
	"winner": {
	"key": "0:1:0:2:1:2:102:14:0:0",
	"fileName": "Model-MQ-Q5_K_1.gguf",
	"displayName": "Model-MQ-Q5_K_1",
	"shortName": "MQ-Q5_K_1",
	"provider": "MagicQuant",
	"quantFamily": "Q5_K",
	"isHybrid": true,
	"isExternalPureBaseline": false,
	"kld": 0.006632,
	"ppl": 8.900069,
	"pplDeltaPercent": 0.1853865550002334,
	"sizeBytes": 3090906816,
	"sizeGiB": 2.878631293773651
	},
	"deltas": {
	"kld": 0.00029299999999999986,
	"sizeBytes": -50135040,
	"sizeGiB": -0.04669189453125,
	"removedPplDeltaPercent": 0.204354090683967,
	"winnerPplDeltaPercent": 0.1853865550002334,
	"pplDeltaPercentImprovement": 0.018967535683733605
	}
	},
	{
	"reasonCode": "SPACING_COLLAPSE",
	"reasonDescription": "Two candidates were too close in practical output space; the stronger one was kept.",
	"rawReason": "meaningful spacing collapse; size gap below 91,013,530 bytes",
	"removed": {
	"key": "2:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-LM-Q5_K.gguf",
	"displayName": "Model-LM-Q5_K",
	"shortName": "LM-Q5_K",
	"provider": "llama.cpp",
	"quantFamily": "Q5_K",
	"isHybrid": false,
	"isExternalPureBaseline": false,
	"kld": 0.009942,
	"ppl": 8.896878,
	"pplDeltaPercent": 0.14946643252735017,
	"sizeBytes": 2889514176,
	"sizeGiB": 2.691069781780243
	},
	"winner": {
	"key": "101:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-UD-Q5_K_XL.gguf",
	"displayName": "Model-UD-Q5_K_XL",
	"shortName": "UD-Q5_K_XL",
	"provider": "Unsloth",
	"quantFamily": "UD-Q5_K_XL",
	"isHybrid": false,
	"isExternalPureBaseline": true,
	"kld": 0.009839,
	"ppl": 8.923979,
	"pplDeltaPercent": 0.45453419784771626,
	"sizeBytes": 2930382016,
	"sizeGiB": 2.7291309237480164
	},
	"deltas": {
	"kld": 0.00010299999999999893,
	"sizeBytes": -40867840,
	"sizeGiB": -0.03806114196777344,
	"removedPplDeltaPercent": 0.14946643252735017,
	"winnerPplDeltaPercent": 0.45453419784771626,
	"pplDeltaPercentImprovement": -0.30506776532036606
	}
	},
	{
	"reasonCode": "SPACING_COLLAPSE",
	"reasonDescription": "Two candidates were too close in practical output space; the stronger one was kept.",
	"rawReason": "meaningful spacing collapse; size gap below 91,013,530 bytes",
	"removed": {
	"key": "13:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-LM-Q5_K_S.gguf",
	"displayName": "Model-LM-Q5_K_S",
	"shortName": "LM-Q5_K_S",
	"provider": "llama.cpp",
	"quantFamily": "Q5_K_S",
	"isHybrid": false,
	"isExternalPureBaseline": false,
	"kld": 0.010808,
	"ppl": 8.919999,
	"pplDeltaPercent": 0.40973254086182603,
	"sizeBytes": 2823711936,
	"sizeGiB": 2.6297866702079773
	},
	"winner": {
	"key": "2:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-LM-Q5_K.gguf",
	"displayName": "Model-LM-Q5_K",
	"shortName": "LM-Q5_K",
	"provider": "llama.cpp",
	"quantFamily": "Q5_K",
	"isHybrid": false,
	"isExternalPureBaseline": false,
	"kld": 0.009942,
	"ppl": 8.896878,
	"pplDeltaPercent": 0.14946643252735017,
	"sizeBytes": 2889514176,
	"sizeGiB": 2.691069781780243
	},
	"deltas": {
	"kld": 0.0008660000000000004,
	"sizeBytes": -65802240,
	"sizeGiB": -0.061283111572265625,
	"removedPplDeltaPercent": 0.40973254086182603,
	"winnerPplDeltaPercent": 0.14946643252735017,
	"pplDeltaPercentImprovement": 0.26026610833447583
	}
	},
	{
	"reasonCode": "SPACING_COLLAPSE",
	"reasonDescription": "Two candidates were too close in practical output space; the stronger one was kept.",
	"rawReason": "meaningful spacing collapse; size gap below 91,013,530 bytes",
	"removed": {
	"key": "0:1:0:7:1:2:7:15:0:0",
	"fileName": "Model-MQ-Q4_K_M.gguf",
	"displayName": "Model-MQ-Q4_K_M",
	"shortName": "MQ-Q4_K_M",
	"provider": "MagicQuant",
	"quantFamily": "Q8_0",
	"isHybrid": true,
	"isExternalPureBaseline": false,
	"kld": 0.021705,
	"ppl": 8.964544,
	"pplDeltaPercent": 0.9111621414741836,
	"sizeBytes": 2600062656,
	"sizeGiB": 2.4214970469474792
	},
	"winner": {
	"key": "0:1:0:3:1:2:7:7:0:0",
	"fileName": "Model-MQ-Q4_K_M_1.gguf",
	"displayName": "Model-MQ-Q4_K_M_1",
	"shortName": "MQ-Q4_K_M_1",
	"provider": "MagicQuant",
	"quantFamily": "Q4_K_M",
	"isHybrid": true,
	"isExternalPureBaseline": false,
	"kld": 0.020346,
	"ppl": 8.95744,
	"pplDeltaPercent": 0.8311945607636608,
	"sizeBytes": 2618576576,
	"sizeGiB": 2.4387394785881042
	},
	"deltas": {
	"kld": 0.001358999999999999,
	"sizeBytes": -18513920,
	"sizeGiB": -0.017242431640625,
	"removedPplDeltaPercent": 0.9111621414741836,
	"winnerPplDeltaPercent": 0.8311945607636608,
	"pplDeltaPercentImprovement": 0.07996758071052279
	}
	},
	{
	"reasonCode": "NEAR_BASELINE_PREMIUM",
	"reasonDescription": "The winner used only the configured near-baseline size premium and beat the real linear KLD trade line.",
	"rawReason": "near-baseline replacement within \u002B1% size premium",
	"removed": {
	"key": "100:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-UD-Q4_K_XL.gguf",
	"displayName": "Model-UD-Q4_K_XL",
	"shortName": "UD-Q4_K_XL",
	"provider": "Unsloth",
	"quantFamily": "UD-Q4_K_XL",
	"isHybrid": false,
	"isExternalPureBaseline": true,
	"kld": 0.022351,
	"ppl": 8.997353,
	"pplDeltaPercent": 1.2804831374668026,
	"sizeBytes": 2591284416,
	"sizeGiB": 2.4133216738700867
	},
	"winner": {
	"key": "0:1:0:7:1:2:7:15:0:0",
	"fileName": "Model-MQ-Q4_K_M.gguf",
	"displayName": "Model-MQ-Q4_K_M",
	"shortName": "MQ-Q4_K_M",
	"provider": "MagicQuant",
	"quantFamily": "Q8_0",
	"isHybrid": true,
	"isExternalPureBaseline": false,
	"kld": 0.021705,
	"ppl": 8.964544,
	"pplDeltaPercent": 0.9111621414741836,
	"sizeBytes": 2600062656,
	"sizeGiB": 2.4214970469474792
	},
	"deltas": {
	"kld": 0.0006460000000000007,
	"sizeBytes": -8778240,
	"sizeGiB": -0.008175373077392578,
	"removedPplDeltaPercent": 1.2804831374668026,
	"winnerPplDeltaPercent": 0.9111621414741836,
	"pplDeltaPercentImprovement": 0.36932099599261903
	}
	},
	{
	"reasonCode": "SPACING_COLLAPSE",
	"reasonDescription": "Two candidates were too close in practical output space; the stronger one was kept.",
	"rawReason": "meaningful spacing collapse; size gap below 91,013,530 bytes",
	"removed": {
	"key": "100:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-UD-Q4_K_XL.gguf",
	"displayName": "Model-UD-Q4_K_XL",
	"shortName": "UD-Q4_K_XL",
	"provider": "Unsloth",
	"quantFamily": "UD-Q4_K_XL",
	"isHybrid": false,
	"isExternalPureBaseline": true,
	"kld": 0.022351,
	"ppl": 8.997353,
	"pplDeltaPercent": 1.2804831374668026,
	"sizeBytes": 2591284416,
	"sizeGiB": 2.4133216738700867
	},
	"winner": {
	"key": "0:1:0:7:1:2:7:15:0:0",
	"fileName": "Model-MQ-Q4_K_M.gguf",
	"displayName": "Model-MQ-Q4_K_M",
	"shortName": "MQ-Q4_K_M",
	"provider": "MagicQuant",
	"quantFamily": "Q8_0",
	"isHybrid": true,
	"isExternalPureBaseline": false,
	"kld": 0.021705,
	"ppl": 8.964544,
	"pplDeltaPercent": 0.9111621414741836,
	"sizeBytes": 2600062656,
	"sizeGiB": 2.4214970469474792
	},
	"deltas": {
	"kld": 0.0006460000000000007,
	"sizeBytes": -8778240,
	"sizeGiB": -0.008175373077392578,
	"removedPplDeltaPercent": 1.2804831374668026,
	"winnerPplDeltaPercent": 0.9111621414741836,
	"pplDeltaPercentImprovement": 0.36932099599261903
	}
	},
	{
	"reasonCode": "SPACING_COLLAPSE",
	"reasonDescription": "Two candidates were too close in practical output space; the stronger one was kept.",
	"rawReason": "meaningful spacing collapse; size gap below 91,013,530 bytes",
	"removed": {
	"key": "0:1:0:7:1:2:7:7:0:0",
	"fileName": "Model-MQ-Q4_K_M.gguf",
	"displayName": "Model-MQ-Q4_K_M",
	"shortName": "MQ-Q4_K_M",
	"provider": "MagicQuant",
	"quantFamily": "Q8_0",
	"isHybrid": true,
	"isExternalPureBaseline": false,
	"kld": 0.023119,
	"ppl": 8.965156,
	"pplDeltaPercent": 0.9180512404881,
	"sizeBytes": 2559594176,
	"sizeGiB": 2.3838078379631042
	},
	"winner": {
	"key": "100:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-UD-Q4_K_XL.gguf",
	"displayName": "Model-UD-Q4_K_XL",
	"shortName": "UD-Q4_K_XL",
	"provider": "Unsloth",
	"quantFamily": "UD-Q4_K_XL",
	"isHybrid": false,
	"isExternalPureBaseline": true,
	"kld": 0.022351,
	"ppl": 8.997353,
	"pplDeltaPercent": 1.2804831374668026,
	"sizeBytes": 2591284416,
	"sizeGiB": 2.4133216738700867
	},
	"deltas": {
	"kld": 0.0007680000000000013,
	"sizeBytes": -31690240,
	"sizeGiB": -0.029513835906982422,
	"removedPplDeltaPercent": 0.9180512404881,
	"winnerPplDeltaPercent": 1.2804831374668026,
	"pplDeltaPercentImprovement": -0.3624318969787026
	}
	},
	{
	"reasonCode": "SPACING_COLLAPSE",
	"reasonDescription": "Two candidates were too close in practical output space; the stronger one was kept.",
	"rawReason": "meaningful spacing collapse; size gap below 91,013,530 bytes",
	"removed": {
	"key": "3:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-LM-Q4_K_M.gguf",
	"displayName": "Model-LM-Q4_K_M",
	"shortName": "LM-Q4_K_M",
	"provider": "llama.cpp",
	"quantFamily": "Q4_K_M",
	"isHybrid": false,
	"isExternalPureBaseline": false,
	"kld": 0.025432,
	"ppl": 9.030426,
	"pplDeltaPercent": 1.6527759016614976,
	"sizeBytes": 2497281216,
	"sizeGiB": 2.325774371623993
	},
	"winner": {
	"key": "0:1:0:7:1:2:7:7:0:0",
	"fileName": "Model-MQ-Q4_K_M.gguf",
	"displayName": "Model-MQ-Q4_K_M",
	"shortName": "MQ-Q4_K_M",
	"provider": "MagicQuant",
	"quantFamily": "Q8_0",
	"isHybrid": true,
	"isExternalPureBaseline": false,
	"kld": 0.023119,
	"ppl": 8.965156,
	"pplDeltaPercent": 0.9180512404881,
	"sizeBytes": 2559594176,
	"sizeGiB": 2.3838078379631042
	},
	"deltas": {
	"kld": 0.002312999999999999,
	"sizeBytes": -62312960,
	"sizeGiB": -0.05803346633911133,
	"removedPplDeltaPercent": 1.6527759016614976,
	"winnerPplDeltaPercent": 0.9180512404881,
	"pplDeltaPercentImprovement": 0.7347246611733975
	}
	},
	{
	"reasonCode": "SPACING_COLLAPSE",
	"reasonDescription": "Two candidates were too close in practical output space; the stronger one was kept.",
	"rawReason": "meaningful spacing collapse; size gap below 91,013,530 bytes",
	"removed": {
	"key": "5:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-LM-IQ4_NL.gguf",
	"displayName": "Model-LM-IQ4_NL",
	"shortName": "LM-IQ4_NL",
	"provider": "llama.cpp",
	"quantFamily": "IQ4_NL",
	"isHybrid": false,
	"isExternalPureBaseline": false,
	"kld": 0.030626,
	"ppl": 8.946429,
	"pplDeltaPercent": 0.7072470620018989,
	"sizeBytes": 2381343936,
	"sizeGiB": 2.2177993655204773
	},
	"winner": {
	"key": "14:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-LM-Q4_K_S.gguf",
	"displayName": "Model-LM-Q4_K_S",
	"shortName": "LM-Q4_K_S",
	"provider": "llama.cpp",
	"quantFamily": "Q4_K_S",
	"isHybrid": false,
	"isExternalPureBaseline": false,
	"kld": 0.029803,
	"ppl": 9.010033,
	"pplDeltaPercent": 1.4232180647485306,
	"sizeBytes": 2383310016,
	"sizeGiB": 2.2196304202079773
	},
	"deltas": {
	"kld": 0.0008230000000000008,
	"sizeBytes": -1966080,
	"sizeGiB": -0.0018310546875,
	"removedPplDeltaPercent": 0.7072470620018989,
	"winnerPplDeltaPercent": 1.4232180647485306,
	"pplDeltaPercentImprovement": -0.7159710027466317
	}
	},
	{
	"reasonCode": "SPACING_COLLAPSE",
	"reasonDescription": "Two candidates were too close in practical output space; the stronger one was kept.",
	"rawReason": "meaningful spacing collapse; size gap below 91,013,530 bytes",
	"removed": {
	"key": "8:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-LM-IQ3_XS.gguf",
	"displayName": "Model-LM-IQ3_XS",
	"shortName": "LM-IQ3_XS",
	"provider": "llama.cpp",
	"quantFamily": "IQ3_XS",
	"isHybrid": false,
	"isExternalPureBaseline": false,
	"kld": 0.11761,
	"ppl": 9.485851,
	"pplDeltaPercent": 6.779357467693287,
	"sizeBytes": 1814375616,
	"sizeGiB": 1.6897689700126648
	},
	"winner": {
	"key": "7:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-LM-IQ3_S.gguf",
	"displayName": "Model-LM-IQ3_S",
	"shortName": "LM-IQ3_S",
	"provider": "llama.cpp",
	"quantFamily": "IQ3_S",
	"isHybrid": false,
	"isExternalPureBaseline": false,
	"kld": 0.091992,
	"ppl": 9.349099,
	"pplDeltaPercent": 5.239981539015728,
	"sizeBytes": 1899531456,
	"sizeGiB": 1.7690765261650085
	},
	"deltas": {
	"kld": 0.025618000000000002,
	"sizeBytes": -85155840,
	"sizeGiB": -0.07930755615234375,
	"removedPplDeltaPercent": 6.779357467693287,
	"winnerPplDeltaPercent": 5.239981539015728,
	"pplDeltaPercentImprovement": 1.5393759286775586
	}
	},
	{
	"reasonCode": "SPACING_COLLAPSE",
	"reasonDescription": "Two candidates were too close in practical output space; the stronger one was kept.",
	"rawReason": "meaningful spacing collapse; size gap below 91,013,530 bytes",
	"removed": {
	"key": "11:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-LM-IQ2_XS.gguf",
	"displayName": "Model-LM-IQ2_XS",
	"shortName": "LM-IQ2_XS",
	"provider": "llama.cpp",
	"quantFamily": "IQ2_XS",
	"isHybrid": false,
	"isExternalPureBaseline": false,
	"kld": 0.529564,
	"ppl": 13.498521,
	"pplDeltaPercent": 51.94877076860733,
	"sizeBytes": 1354100416,
	"sizeGiB": 1.2611042857170105
	},
	"winner": {
	"key": "10:0:0:0:0:0:0:0:0:0",
	"fileName": "Model-LM-IQ2_S.gguf",
	"displayName": "Model-LM-IQ2_S",
	"shortName": "LM-IQ2_S",
	"provider": "llama.cpp",
	"quantFamily": "IQ2_S",
	"isHybrid": false,
	"isExternalPureBaseline": false,
	"kld": 0.431128,
	"ppl": 12.334448,
	"pplDeltaPercent": 38.84515286595525,
	"sizeBytes": 1417301696,
	"sizeGiB": 1.3199650645256042
	},
	"deltas": {
	"kld": 0.09843600000000002,
	"sizeBytes": -63201280,
	"sizeGiB": -0.05886077880859375,
	"removedPplDeltaPercent": 51.94877076860733,
	"winnerPplDeltaPercent": 38.84515286595525,
	"pplDeltaPercentImprovement": 13.103617902652083
	}
	}
	]