EXAONE 4.0: Unified Large Language Models Integrating Non-reasoning and Reasoning Modes

TL;DR

EXAONE 4.0 is a unified large language model that combines non-reasoning and reasoning modes, supports multiple languages, and is optimized for both high performance and on-device use, advancing agentic AI capabilities.

Contribution

The paper introduces EXAONE 4.0, a novel large language model integrating reasoning and non-reasoning modes with multilingual support and agentic features, available in multiple sizes.

Findings

EXAONE 4.0 outperforms comparable open-weight models.

It maintains competitiveness with frontier-class models.

Models are publicly accessible for research.

Abstract

This technical report introduces EXAONE 4.0, which integrates a Non-reasoning mode and a 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. The EXAONE 4.0 demonstrates superior performance compared to open-weight models in its class and remains competitive even against frontier-class models. The models are publicly available for research purposes and can be easily downloaded via https://huggingface.co/LGAI-EXAONE.