Naver’s cloud unit, Naver Cloud, which launched the HyperCLOVA X, said CLOVA X started its beta service in English and Korean on August 24. Cue, available in September for beta testing, will be incorporated into the current Naver’s search engine in November.

HyperCLOVA X is available for creators and enterprise customers. HyperCLOVA X is an upgraded version of its previous LLM for the Korean language, called HyperCLOVA, launched in 2021. Naver says HyperCLOVA has more than 204 billion parameters, but it did not disclose how many parameters have been trained on the HyperCLOVA X. (Users must sign up for Naver accounts to use HyperCLOVA X, the company said.)

Naver has more than 500 AI experts on its team and is one of the five companies worldwide that has developed its own large language model with more than 100 billion parameters, CEO of Naver SooYeon Choi said in a letter to shareholders published earlier this week.

Naver Cloud is also developing CLOVA X’s multimodal language models that will be capable of analyzing and generating not only text but images, videos and audio, said Nako Sung, Head of Technology and Hypersacale AI at Naver Cloud, said at the company’s conference on Thursday. Naver aims to release the Clova X multimodal language model in the near future, according to Sung.

Naver’s generative AI rollout comes as global tech players try to stay ahead of enhancing their own generative AI products. Microsoft-backed OpenAI has sparked competition in the generative AI space. Google launched its AI chatbot Bard and invested in Anthropic and Baidu released Ernie Bot. Companies like Meta and Amazon also are reportedly preparing to release their own AI-powered chatbots.

Naver said it has been looking to the regions the competitors have not yet covered to offer tailored AI applications for countries with political sensitivities in the Middle East and non-English speaking countries like Spain and Mexico in addition to South Korea, Japan and Southeast Asia.

The Korean internet giant will open its second data center, GAK Sejong, with 600,000 servers in South Korea for its generative AI services in November, Choi said at the conference. Naver has been working with Samsung to develop AI chips for hyperscale AI since last December.