South Korea has made a significant stride in the global artificial intelligence (AI) landscape with the launch of its first open-source AI model, EXAONE 3.0, by LG AI Research.
EXAONE 3.0 stands out as a powerful tool, supporting both English and Korean languages. It is built on a Decoder-only Transformer architecture, which allows it to process large amounts of information effectively. The model is equipped with an impressive 7.8 billion parameters and has been trained on a massive dataset of 8 trillion tokens, making it highly capable in understanding and generating human-like text.
EXAONE Benchmark
One of the most notable features is its open-source nature. By making this model publicly available, LG AI Research aims to boost AI development and innovation, not just in South Korea but globally.
EXAONE Model
The model has already demonstrated exceptional performance in various tasks, including English language processing, mathematics, and coding. It also offers significant improvements in terms of faster processing times, lower memory usage, and reduced operating costs, making it more efficient for users.
Looking ahead, LG AI Research plans to expand the training data for EXAONE 3.0 to 100 million cases by the end of the year. This expansion is expected to further enhance the model’s capabilities, allowing it to tackle even more complex challenges in AI.
With the release of EXAONE 3.0, South Korea is taking a bold step into the global AI arena, showcasing its commitment to advancing technology and fostering innovation through open collaboration.