LG AI Research Releases EXAONE Model Series
LG AI Research has introduced the EXAONE family of Large Language Models (LLMs), with multiple versions addressing different capabilities.
EXAONE 3.0
According to the arXiv paper (arXiv:2408.03541v4), EXAONE 3.0 represents “the first open model in the family of Large Language Models (LLMs) developed by LG AI Research.” The company has publicly released a 7.8B instruction-tuned model from this series.
EXAONE Deep
The EXAONE Deep series, detailed in arXiv:2503.12524v3, “exhibits superior capabilities in various reasoning tasks, including math and coding benchmarks.” According to the paper, the models were trained “mainly on the reasoning-specialized dataset that incorporates long streams of thought.”
EXAONE 4.0
The latest iteration, EXAONE 4.0 (arXiv:2507.11407v2), “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.” This unified approach appears designed to combine practical usability with enhanced reasoning capabilities in a single model.
The releases represent LG AI Research’s progression in developing both open and capable language models with specialized reasoning abilities.