Instructions to use second-state/EXAONE-3.5-32B-Instruct-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use second-state/EXAONE-3.5-32B-Instruct-GGUF with Transformers:

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

pipe = pipeline("text-generation", model="second-state/EXAONE-3.5-32B-Instruct-GGUF", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("second-state/EXAONE-3.5-32B-Instruct-GGUF", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use second-state/EXAONE-3.5-32B-Instruct-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 second-state/EXAONE-3.5-32B-Instruct-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf second-state/EXAONE-3.5-32B-Instruct-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 second-state/EXAONE-3.5-32B-Instruct-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf second-state/EXAONE-3.5-32B-Instruct-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 second-state/EXAONE-3.5-32B-Instruct-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf second-state/EXAONE-3.5-32B-Instruct-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 second-state/EXAONE-3.5-32B-Instruct-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf second-state/EXAONE-3.5-32B-Instruct-GGUF:Q4_K_M

Use Docker

docker model run hf.co/second-state/EXAONE-3.5-32B-Instruct-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use second-state/EXAONE-3.5-32B-Instruct-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "second-state/EXAONE-3.5-32B-Instruct-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": "second-state/EXAONE-3.5-32B-Instruct-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/second-state/EXAONE-3.5-32B-Instruct-GGUF:Q4_K_M

SGLang

How to use second-state/EXAONE-3.5-32B-Instruct-GGUF 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 "second-state/EXAONE-3.5-32B-Instruct-GGUF" \
    --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": "second-state/EXAONE-3.5-32B-Instruct-GGUF",
		"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 "second-state/EXAONE-3.5-32B-Instruct-GGUF" \
        --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": "second-state/EXAONE-3.5-32B-Instruct-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use second-state/EXAONE-3.5-32B-Instruct-GGUF with Ollama:
```
ollama run hf.co/second-state/EXAONE-3.5-32B-Instruct-GGUF:Q4_K_M
```
Unsloth Desktop
Docker Model Runner
How to use second-state/EXAONE-3.5-32B-Instruct-GGUF with Docker Model Runner:
```
docker model run hf.co/second-state/EXAONE-3.5-32B-Instruct-GGUF:Q4_K_M
```

Lemonade

How to use second-state/EXAONE-3.5-32B-Instruct-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull second-state/EXAONE-3.5-32B-Instruct-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.EXAONE-3.5-32B-Instruct-GGUF-Q4_K_M

List all available models

lemonade list

Atomic Chat

EXAONE-3.5-32B-Instruct-GGUF

Original Model

LGAI-EXAONE/EXAONE-3.5-32B-Instruct

Run with LlamaEdge

LlamaEdge version: coming soon

Prompt template

Prompt type: exaone-chat

Prompt string

[|system|]{system_message}[|endofturn|]
[|user|]{user_message_1}
[|assistant|]{assistant_message_1}[|endofturn|]
[|user|]{user_message_2}
[|assistant|]

Context size: 32000

Run as LlamaEdge service

wasmedge --dir .:. --nn-preload default:GGML:AUTO:EXAONE-3.5-32B-Instruct-Q5_K_M.gguf \
  llama-api-server.wasm \
  --prompt-template exaone-chat \
  --ctx-size 32000 \
  --model-name EXAONE-3.5-32B-Instruct

Run as LlamaEdge command app

wasmedge --dir .:. --nn-preload default:GGML:AUTO:EXAONE-3.5-32B-Instruct-Q5_K_M.gguf \
  llama-chat.wasm \
  --prompt-template exaone-chat \
  --ctx-size 32000

Quantized GGUF Models

Name	Quant method	Bits	Size	Use case
EXAONE-3.5-32B-Instruct-Q2_K.gguf	Q2_K	2	11.9 GB	smallest, significant quality loss - not recommended for most purposes
EXAONE-3.5-32B-Instruct-Q3_K_L.gguf	Q3_K_L	3	16.8 GB	small, substantial quality loss
EXAONE-3.5-32B-Instruct-Q3_K_M.gguf	Q3_K_M	3	15.5 GB	very small, high quality loss
EXAONE-3.5-32B-Instruct-Q3_K_S.gguf	Q3_K_S	3	14.0 GB	very small, high quality loss
EXAONE-3.5-32B-Instruct-Q4_0.gguf	Q4_0	4	18.1 GB	legacy; small, very high quality loss - prefer using Q3_K_M
EXAONE-3.5-32B-Instruct-Q4_K_M.gguf	Q4_K_M	4	19.3 GB	medium, balanced quality - recommended
EXAONE-3.5-32B-Instruct-Q4_K_S.gguf	Q4_K_S	4	18.3 GB	small, greater quality loss
EXAONE-3.5-32B-Instruct-Q5_0.gguf	Q5_0	5	22.1 GB	legacy; medium, balanced quality - prefer using Q4_K_M
EXAONE-3.5-32B-Instruct-Q5_K_M.gguf	Q5_K_M	5	22.7 GB	large, very low quality loss - recommended
EXAONE-3.5-32B-Instruct-Q5_K_S.gguf	Q5_K_S	5	22.1 GB	large, low quality loss - recommended
EXAONE-3.5-32B-Instruct-Q6_K.gguf	Q6_K	6	26.3 GB	very large, extremely low quality loss
EXAONE-3.5-32B-Instruct-Q8_0.gguf	Q8_0	8	34.0 GB	very large, extremely low quality loss - not recommended
EXAONE-3.5-32B-Instruct-f16-00001-of-00003.gguf	f16	16	29.8 GB
EXAONE-3.5-32B-Instruct-f16-00002-of-00003.gguf	f16	16	30.0 GB
EXAONE-3.5-32B-Instruct-f16-00003-of-00003.gguf	f16	16	4.23 GB

Quantized with llama.cpp b4932.

Downloads last month: 141

GGUF

Model size

32B params

Architecture

exaone

Hardware compatibility

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

16-bit

Model tree for second-state/EXAONE-3.5-32B-Instruct-GGUF

Base model

LGAI-EXAONE/EXAONE-3.5-32B-Instruct

Quantized

(14)

this model

Collection including second-state/EXAONE-3.5-32B-Instruct-GGUF

EXAONE-3.5 GGUF Models

Collection

LlamaEdge compatible quants for EXAONE-3.5 models. • 3 items • Updated Mar 21, 2025