Instructions to use haanjack/K-EXAONE-236B-A23B-MXFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use haanjack/K-EXAONE-236B-A23B-MXFP4 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="haanjack/K-EXAONE-236B-A23B-MXFP4")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("haanjack/K-EXAONE-236B-A23B-MXFP4")
model = AutoModelForCausalLM.from_pretrained("haanjack/K-EXAONE-236B-A23B-MXFP4")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use haanjack/K-EXAONE-236B-A23B-MXFP4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "haanjack/K-EXAONE-236B-A23B-MXFP4"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "haanjack/K-EXAONE-236B-A23B-MXFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/haanjack/K-EXAONE-236B-A23B-MXFP4

SGLang

How to use haanjack/K-EXAONE-236B-A23B-MXFP4 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "haanjack/K-EXAONE-236B-A23B-MXFP4" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "haanjack/K-EXAONE-236B-A23B-MXFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "haanjack/K-EXAONE-236B-A23B-MXFP4" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "haanjack/K-EXAONE-236B-A23B-MXFP4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use haanjack/K-EXAONE-236B-A23B-MXFP4 with Docker Model Runner:
```
docker model run hf.co/haanjack/K-EXAONE-236B-A23B-MXFP4
```

K-EXAONE-236B-A23B-MXFP4

MXFP4 quantized version of LGAI-EXAONE/K-EXAONE-236B-A23B, with MoE routed expert weights and activations quantized to OCP MXFP4 format.

K-EXAONE-236B-A23B is a Mixture-of-Experts language model developed by LG AI Research, featuring 236B total parameters with 23B active parameters per token, 128 routed experts, and 1 shared expert per layer.

Quantization Details

Property	Value
Base Model	LGAI-EXAONE/K-EXAONE-236B-A23B (236B params, MoE)
Precision	MXFP4 (E2M1 weights + E8M0 scales, group_size=32)
Activation Quantization	MXFP4 (E2M1), Dynamic, group_size=32
Quantization Tool	quanto + AMD Quark 0.11.1 (file-to-file)
Algorithm	RTN (Round-To-Nearest)
Original Size	443 GB
Quantized Size	130 GB
Compression Ratio	3.4x

Excluded Layers

This model uses the attn-excl strategy: all attention projections and all shared expert projections are excluded from quantization, leaving only MoE routed expert weights quantized to MXFP4.

Self-Attention (all 48 layers): q_proj, k_proj, v_proj, o_proj
Shared Expert MLP (all 48 layers): gate_proj, up_proj, down_proj
MoE Router Gates: All mlp.gate layers
Standard exclusions: lm_head, embed_tokens, all norm layers

Total: 382 excluded layers.

Hardware

Quantized on: AMD MI355 (gfx950), 288 GB VRAM
Tested with: vLLM v0.18.2 (ROCm), TP=1
Compatible with: vLLM with Quark quantization support (quant_method: "quark")

Usage

vLLM

vllm serve haanjack/K-EXAONE-236B-A23B-MXFP4 \
  --trust-remote-code \
  --tensor-parallel-size 1 \
  --max-model-len 4096

Quantization Reproduction

from quanto import UnifiedQuantizer, UnifiedConfig

NUM_LAYERS = 48
exclude = ["lm_head", "*embed*", "*norm*", "*.gate"]
for i in range(NUM_LAYERS):
    for proj in ["q_proj", "k_proj", "v_proj", "o_proj"]:
        exclude.append(f"model.layers.{i}.self_attn.{proj}")
    for proj in ["gate_proj", "up_proj", "down_proj"]:
        exclude.append(f"model.layers.{i}.mlp.shared_experts.{proj}")

config = UnifiedConfig(
    model_path="LGAI-EXAONE/K-EXAONE-236B-A23B",
    output_dir="./K-EXAONE-236B-A23B-MXFP4",
    precision="mxfp4",
    sensitivity_analysis=False,
    exclude_layers=exclude,
    trust_remote_code=True,
)
UnifiedQuantizer(config).run()

Credits

Base Model: LG AI Research — K-EXAONE
Quantization: quanto with AMD Quark
Hardware: AMD MI355 (gfx950), 288 GB VRAM

License

This model inherits the EXAONE License from the base model.

Downloads last month: 52

Safetensors

Model size

130B params

Tensor type

F32

BF16

Model tree for haanjack/K-EXAONE-236B-A23B-MXFP4

Base model

LGAI-EXAONE/K-EXAONE-236B-A23B

Quantized

(10)

this model