Method: AWQ (AutoAWQ)
Bits: 4-bit
Group Size: 128
Zero Point: True
Calibration Dataset: Custom high-fidelity dataset including Tool-use, Long Context (Ko-LongRAG), Reasoning, and code generation.

Usage

You can use this model with transformers or vLLM that support AWQ quantization for the exaone4 architecture.

Transformers

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "lesj0610/EXAONE-4.0.1-32B-AWQ"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")

vLLM

vllm serve lesj0610/EXAONE-4.0.1-32B-AWQ --quantization awq

Model Architecture

The EXAONE 4.0 architecture features Hybrid Attention and QK-Reorder-Norm. This quantization was specifically patched to handle the lack of input layernorm in the exaone4 implementation.

✈️ Try on FriendliAI (licensed under commercial purposes)

📢 EXAONE 4.0 is officially supported by HuggingFace transformers! Please check out the guide below

EXAONE-4.0.1-32B

The version 4.0.1 is a patch version to reduce unintended or inappropriate responses.

Introduction

We introduce EXAONE 4.0, which integrates a Non-reasoning mode and Reasoning mode to achieve both the excellent usability of EXAONE 3.5 and the advanced reasoning abilities of EXAONE Deep. To pave the way for the agentic AI era, EXAONE 4.0 incorporates essential features such as agentic tool use, and its multilingual capabilities are extended to support Spanish in addition to English and Korean.

The EXAONE 4.0 model series consists of two sizes: a mid-size 32B model optimized for high performance, and a small-size 1.2B model designed for on-device applications.

In the EXAONE 4.0 architecture, we apply new architectural changes compared to previous EXAONE models as below:

Hybrid Attention: For the 32B model, we adopt hybrid attention scheme, which combines Local attention (sliding window attention) with Global attention (full attention) in a 3:1 ratio. We do not use RoPE (Rotary Positional Embedding) for global attention for better global context understanding.
QK-Reorder-Norm: We reorder the LayerNorm position from the traditional Pre-LN scheme by applying LayerNorm directly to the attention and MLP outputs, and we add RMS normalization right after the Q and K projection. It helps yield better performance on downstream tasks despite consuming more computation.

For more details, please refer to our technical report, HuggingFace paper, blog, and GitHub.

Model Configuration

Number of Parameters (without embeddings): 30.95B
Number of Layers: 64
Number of Attention Heads: GQA with 40-heads and 8-KV heads
Vocab Size: 102,400
Context Length: 131,072 tokens

Quickstart

You should install the transformers library with version >= 4.54.0.

Non-reasoning mode

For general use, you can use the EXAONE 4.0 models with the following example:

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "LGAI-EXAONE/EXAONE-4.0.1-32B"

model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="bfloat16",
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained(model_name)

# choose your prompt
prompt = "Explain how wonderful you are"
prompt = "Explica lo increíble que eres"
prompt = "너가 얼마나 대단한지 설명해 봐"

messages = [
    {"role": "user", "content": prompt}
]
input_ids = tokenizer.apply_chat_template(
    messages,
    tokenize=True,
    add_generation_prompt=True,
    return_tensors="pt"
)

output = model.generate(
    input_ids.to(model.device),
    max_new_tokens=128,
    do_sample=False,
)
print(tokenizer.decode(output[0]))

Reasoning mode

The EXAONE 4.0 models have reasoning capabilities for handling complex problems. You can activate reasoning mode by using the enable_thinking=True argument with the tokenizer, which opens a reasoning block that starts with <think> tag without closing it.

python

messages = [
    {"role": "user", "content": "Which one is bigger, 3.12 vs 3.9?"}
]
input_ids = tokenizer.apply_chat_template(
    messages,
    tokenize=True,
    add_generation_prompt=True,
    return_tensors="pt",
    enable_thinking=True,
)

output = model.generate(
    input_ids.to(model.device),
    max_new_tokens=128,
    do_sample=True,
    temperature=0.6,
    top_p=0.95
)
print(tokenizer.decode(output[0]))

[!IMPORTANT] The model generation with reasoning mode can be affected sensitively by sampling parameters, so please refer to the Usage Guideline for better quality.

Agentic tool use

The EXAONE 4.0 models can be used as agents with their tool calling capabilities. You can provide tool schemas to the model for effective tool calling.

python

import random

def roll_dice(max_num: int):
    return random.randint(1, max_num)

tools = [
    {
        "type": "function",
        "function": {
            "name": "roll_dice",
            "description": "Roll a dice with the number 1 to N. User can select the number N.",
            "parameters": {
                "type": "object",
                "required": ["max_num"],
                "properties": {
                    "max_num": {
                        "type": "int",
                        "description": "Max number of the dice"
                    }
                }
            }
        }
    }
]

messages = [
    {"role": "user", "content": "Roll D6 dice twice!"}
]
input_ids = tokenizer.apply_chat_template(
    messages,
    tokenize=True,
    add_generation_prompt=True,
    return_tensors="pt",
    tools=tools,
)

output = model.generate(
    input_ids.to(model.device),
    max_new_tokens=1024,
    do_sample=True,
    temperature=0.6,
    top_p=0.95,
)
print(tokenizer.decode(output[0]))

Deployment

TensorRT-LLM

TensorRT-LLM officially supports EXAONE 4.0 models in the latest commits. Before it is released, you need to clone the TensorRT-LLM repository to build from source.

git clone https://github.com/NVIDIA/TensorRT-LLM.git

After cloning the repository, you need to build the source for installation. Please refer to the official documentation for a guide to build the TensorRT-LLM environment.

You can run the TensorRT-LLM server by following steps:

Write extra configuration YAML file

yaml
1# extra_llm_api_config.yaml 2kv_cache_config: 3 enable_block_reuse: false
Run server with the configuration

trtllm-serve serve LGAI-EXAONE/EXAONE-4.0.1-32B --backend pytorch --extra_llm_api_options extra_llm_api_config.yaml

For more details, please refer to the documentation of EXAONE from TensorRT-LLM.

vLLM

vLLM officially supports EXAONE 4.0 models in the version of 0.10.0. You can run the vLLM server by following command:

vllm serve LGAI-EXAONE/EXAONE-4.0.1-32B --enable-auto-tool-choice --tool-call-parser hermes --reasoning-parser deepseek_r1

For more details, please refer to the vLLM documentation.

[!NOTE] Other inference engines including sglang don't support the EXAONE 4.0 officially now. We will update as soon as these libraries are updated.

Performance

The following tables show the evaluation results of each model, with reasoning and non-reasoning mode. The evaluation details can be found in the technical report.

✅ denotes the model has a hybrid reasoning capability, evaluated by selecting reasoning / non-reasoning on the purpose.
To assess Korean practical and professional knowledge, we adopt both the KMMLU-Redux and KMMLU-Pro benchmarks. Both datasets are publicly released!

32B Reasoning Mode

	EXAONE 4.0.1 32B	Phi 4 reasoning-plus	Magistral Small-2506	Qwen 3 32B	Qwen 3 235B	DeepSeek R1-0528
Model Size	32.0B	14.7B	23.6B	32.8B	235B	671B
Hybrid Reasoning	✅			✅	✅
World Knowledge
MMLU-Pro	81.8	76.0	73.4	80.0	83.0	85.0
GPQA-Diamond	74.3	68.9	68.2	68.4	71.1	81.0
Math/Coding
AIME 2025	84.5	78.0	62.8	72.9	81.5	87.5
LiveCodeBench v6	67.7	47.1	47.4	60.1	58.9	70.3
Instruction Following
IFEval	82.3	84.9	37.9	85.0	83.4	80.8
Agentic Tool Use
BFCL-v3	60.7	N/A	40.4	70.3	70.8	64.7
Tau-Bench (Airline)	48.0	N/A	38.5	34.5	37.5	53.5
Tau-Bench (Retail)	65.4	N/A	10.2	55.2	58.3	63.9
Multilinguality
KMMLU-Pro	65.7	55.8	51.5	61.4	68.1	71.7
KSM	87.0	79.8	71.9	82.8	86.2	86.7
MMMLU (ES)	85.4	84.3	68.9	82.8	86.7	88.2

32B Non-Reasoning Mode

	EXAONE 4.0.1 32B	Phi 4	Mistral-Small-2506	Gemma3 27B	Qwen3 32B	Qwen3 235B	Llama-4-Maverick	DeepSeek V3-0324
Model Size	32.0B	14.7B	24.0B	27.4B	32.8B	235B	402B	671B
Hybrid Reasoning	✅				✅	✅
World Knowledge
MMLU-Pro	77.4	70.4	69.1	67.5	74.4	77.4	80.5	81.2
GPQA-Diamond	61.6	56.1	46.1	42.4	54.6	62.9	69.8	68.4
Math/Coding
AIME 2025	36.3	17.8	30.2	23.8	20.2	24.7	18.0	50.0
LiveCodeBench v6	43.3	27.4	26.9	29.7	28.0	31.4	32.7	44.0
Instruction Following
IFEval	84.7	63.0	77.8	82.6	83.2	83.2	85.4	81.2
Agentic Tool Use
BFCL-v3	63.9	N/A	57.7	N/A	63.0	68.0	52.9	63.8
Tau-Bench (Airline)	18.5	N/A	36.1	N/A	16.0	27.0	38.0	40.5
Tau-Bench (Retail)	52.0	N/A	35.5	N/A	47.6	56.5	6.5	68.5
Multilinguality
KMMLU-Pro	59.8	44.8	51.0	50.7	58.3	64.4	68.8	67.3
KSM	56.3	29.1	35.5	36.1	41.3	46.6	40.6	63.5
MMMLU (ES)	80.3	81.2	78.4	78.7	82.1	83.7	86.9	86.7

Usage Guideline

[!IMPORTANT] To achieve the expected performance, we recommend using the following configurations:

For non-reasoning mode, we recommend using a lower temperature value such as temperature<0.6 for better performance.

For reasoning mode (using <think> block), we recommend using temperature=0.6 and top_p=0.95.

If you suffer from the model degeneration, we recommend using presence_penalty=1.5.

For Korean general conversation with 1.2B model, we suggest to use temperature=0.1 to avoid code switching.

Limitation

The EXAONE language model has certain limitations and may occasionally generate inappropriate responses. The language model generates responses based on the output probability of tokens, and it is determined during learning from training data. While we have made every effort to exclude personal, harmful, and biased information from the training data, some problematic content may still be included, potentially leading to undesirable responses. Please note that the text generated by EXAONE language model does not reflect the views of LG AI Research.

Inappropriate answers may be generated, which contain personal, harmful or other inappropriate information.
Biased responses may be generated, which are associated with age, gender, race, and so on.
The generated responses rely heavily on statistics from the training data, which can result in the generation of semantically or syntactically incorrect sentences.
Since the model does not reflect the latest information, the responses may be false or contradictory.

LG AI Research strives to reduce potential risks that may arise from EXAONE language models. Users are not allowed to engage in any malicious activities (e.g., keying in illegal information) that may induce the creation of inappropriate outputs violating LG AI's ethical principles when using EXAONE language models.

License

The model is licensed under EXAONE AI Model License Agreement 1.2 - NC

[!NOTE] The main difference from the older version is as below:

We removed the claim of model output ownership from the license.

We restrict the model use against the development of models that compete with EXAONE.

We allow the model to be used for educational purposes, not just research.

Citation

@article{exaone-4.0,
  title={EXAONE 4.0: Unified Large Language Models Integrating Non-reasoning and Reasoning Modes},
  author={{LG AI Research}},
  journal={arXiv preprint arXiv:2507.11407},
  year={2025}
}

Contact

LG AI Research Technical Support: contact_us@lgresearch.ai