Naver, the prominent internet search company in Korea, has introduced its expansive AI model called “HyperCLOVA X,” specifically designed to cater to the requirements of the Korean populace. The company believes the AI transformation from the new generative AI model will spark a new era for business.
Leveraging the extensive dataset accumulated over the 24 years through its search engine, Naver envisions that the new AI model can enhance the user experience when pitted against other substantial language models released by Microsoft and Alphabet’s Google, according to Naver Chief Executive Officer Choi Soo-yeon, in a conference held in Seoul on Thursday.
While Naver has not disclosed the specifics of the recently launched HyperClova X, the parameters for its previous version revealed by the search giant two years ago totaled 204 billion. This figure surpasses OpenAI’s ChatGPT 3.5, which stands at 175 billion parameters.
Naver has refrained from disclosing its sales and operating profit goals for the HyperClova X, focusing instead on its potential impact within the B2B sector. To cater to this market, Naver introduced the “NeuroCloud for HyperClova X,” a hybrid cloud solution designed for enterprises.
The product involves the installation of NeuroCloud directly into the client’s data center, coupled with a graphics processing unit cluster to power the HyperClova X model. By establishing a closed network environment within the confines of the data center, Naver argues NeuroCloud enables the creation of a customized, language model without the risk of compromising security or data leakage.