Instructions to use Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF",
	filename="Kimi-Linear-48B-A3B-Instruct-APEX-balanced.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 Myric/Kimi-Linear-48B-A3B-Instruct-APEX-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 Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF
# Run inference directly in the terminal:
llama cli -hf Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF
# Run inference directly in the terminal:
llama cli -hf Myric/Kimi-Linear-48B-A3B-Instruct-APEX-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 Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF
# Run inference directly in the terminal:
./llama-cli -hf Myric/Kimi-Linear-48B-A3B-Instruct-APEX-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 Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF

Use Docker

docker model run hf.co/Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF

LM Studio
Jan

vLLM

How to use Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Myric/Kimi-Linear-48B-A3B-Instruct-APEX-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": "Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF

Ollama
How to use Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF with Ollama:
```
ollama run hf.co/Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF
```

Unsloth Studio

How to use Myric/Kimi-Linear-48B-A3B-Instruct-APEX-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 Myric/Kimi-Linear-48B-A3B-Instruct-APEX-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 Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF to start chatting

How to use Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Myric/Kimi-Linear-48B-A3B-Instruct-APEX-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": "Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use Myric/Kimi-Linear-48B-A3B-Instruct-APEX-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 Myric/Kimi-Linear-48B-A3B-Instruct-APEX-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 Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF

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 "Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF" \
  --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 Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF with Docker Model Runner:
```
docker model run hf.co/Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF
```

Lemonade

How to use Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF

Run and chat with the model

lemonade run user.Kimi-Linear-48B-A3B-Instruct-APEX-GGUF-{{QUANT_TAG}}

List all available models

lemonade list

Kimi-Linear-48B-A3B-Instruct — APEX GGUF

MoE-aware, mixed-precision APEX quantizations of moonshotai/Kimi-Linear-48B-A3B-Instruct — 48B total / ~3B active, a hybrid linear-attention MoE: most layers use KDA (Kimi Delta Attention, gated-delta linear attention), a few use full MLA attention, over a 256-routed + 1-shared expert FFN.

To my knowledge this is the first APEX quant of a linear-attention hybrid MoE. APEX assigns precision per tensor role and per layer instead of uniformly; here that meant teaching the recipe about tensor families the stock generator doesn't know (see Method).

Results

Perplexity on wikitext-2-raw (test, 200×512-token windows), llama-perplexity.

File	Size	BPW	PPL	Δ vs bf16
bf16 (reference)	92 GB	16.0	7.374	—
APEX-balanced (no-imatrix)	33 GB	5.72	7.377	+0.04%
APEX-handroll (ssm@Q8_0, no-imatrix)	33 GB	5.72	7.382	+0.11%
APEX-i-quality (imatrix, IQ4_XS mid experts)	29 GB	4.94	pending	pending

Both no-imatrix tiers land essentially on the bf16 reference (within ~0.1%). An imatrix-guided i-quality tier (IQ4_XS mid experts, 29 GB) is now available as Kimi-Linear-48B-A3B-Instruct-APEX-i-quality.gguf; its perplexity benchmark is still being run and will be filled in here.

From a 92 GB bf16 baseline → 33 GB (~2.8× smaller), and it runs on a 128 GB unified-memory box (fits with full GPU offload). Coherent on general and factual prompts.

The balanced and handroll tiers were built without an imatrix (Q6_K/Q5_K experts, Q8_0 shared, Q6_K attention — none of which require importance data). The newer i-quality tier is imatrix-guided (IQ4_XS mid experts). Deeper lower-bit "I-tier" variants (IQ3/IQ2) would also need an imatrix and are not included here.

Note on the two tiers (a null result)

The hand-roll tier pins the KDA recurrence tensors (ssm_conv1d_*, ssm_f/g_*, ssm_beta) to Q8_0 instead of Q6_K, testing whether protecting the linear-attention state preserves quality. It doesn't — PPL is identical within noise (7.382 vs 7.377), at the same size (the ssm tensors are tiny next to the experts). Use balanced. The hand-roll is kept only to document the experiment.

Which file

APEX-balanced — recommended. Q6_K/Q5_K experts on a layer-depth gradient, Q8_0 shared experts, Q6_K attention + KDA tensors.
APEX-i-quality — imatrix-guided IQ4_XS mid experts (29 GB); the smallest tier here. Try it when you want to save a few GB over balanced; PPL comparison pending.
APEX-handroll — experimental (see null-result note below); not recommended.

Usage (llama.cpp)

llama-cli   -m Kimi-Linear-48B-A3B-Instruct-APEX-balanced.gguf -ngl 999 -p "Hello"
llama-server -m Kimi-Linear-48B-A3B-Instruct-APEX-balanced.gguf -ngl 999 --host 0.0.0.0 --port 8080

Requires a llama.cpp build supporting the kimi_linear architecture and the kimi-k2 pre-tokenizer.

Method

APEX is a bit-allocation recipe over stock llama-quantize --tensor-type-file. Kimi-Linear needed two tensor families the stock APEX generator doesn't emit:

MLA (full-attention layers): attn_kv_a_mqa, attn_k_b, attn_v_b
KDA (linear-attention layers): ssm_conv1d_{k,q,v}, ssm_f_a/f_b, ssm_g_a/g_b, ssm_beta (norms/1-D state kept F32)

plus a dense layer 0 (--dense-layers 1). The expert intermediate dim is 2048 (256-divisible), so no IQ4_NL workaround was needed. Config generation + patching: see REPRODUCE.md, patch_kimi_config.py, and configs/.

Baseline: quantized from bartowski's bf16 GGUF.

Attribution & licenses

All MIT; see LICENSE and NOTICE.

Base: Moonshot AI (@moonshotai) — Kimi-Linear-48B-A3B-Instruct (MIT)
bf16 GGUF: bartowski (@bartowski) — source
Engine: llama.cpp (@ggml-org · github) (MIT)
APEX: Ettore Di Giacinto / LocalAI (@mudler) — localai-org/apex-quant (MIT)

Unofficial community quantization; not affiliated with or endorsed by Moonshot AI.

Downloads last month: 690

GGUF

Model size

49B params

Architecture

kimi-linear

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Model tree for Myric/Kimi-Linear-48B-A3B-Instruct-APEX-GGUF

Base model

moonshotai/Kimi-Linear-48B-A3B-Instruct

Quantized

(24)

this model