{"id":3204,"date":"2026-07-29T10:02:03","date_gmt":"2026-07-29T01:02:03","guid":{"rendered":"https:\/\/news.sktelecom.com\/en\/?p=3204"},"modified":"2026-07-29T10:02:03","modified_gmt":"2026-07-29T01:02:03","slug":"sk-telecom-unveils-a-x-k2driving-ai-adoption-in-industry-and-daily-life","status":"publish","type":"post","link":"https:\/\/news.sktelecom.com\/en\/3204","title":{"rendered":"SK Telecom Unveils A.X K2,Driving AI Adoption in Industry and Daily Life"},"content":{"rendered":"

\u00b7\u00a0 Unveils 688-billion-parameter model A.X K2 alongside three lightweight models for advanced multimodal capabilities<\/strong>
\n\u00b7\u00a0 Expands AI across manufacturing, defense and bio while supporting the wider AI ecosystem, including startups<\/strong>
\n\u00b7\u00a0 Develops vision and audio derivative models in parallel to expand manufacturing AI and service adoption<\/strong><\/p>\n

SK Telecom (NYSE: SKM, hereinafter referred to as \u201cSKT\u201d) today unveiled its proprietary AI foundation model \u2018A.X K2,\u2019 supporting AI transformation across everyday life and industrial fields.<\/p>\n

Starting today, the 688-billion-parameter A.X K2<\/span><\/a> model is available through the open-source community Hugging Face. Building on the 519B A.X K1, it delivers enhanced mathematical and scientific reasoning, Korean-language knowledge, and long-context reasoning capabilities.<\/p>\n

Across 14 benchmarks, A.X K2 posted an average performance gain of 32.2 percentage points over A.X K1. The model delivered particularly strong gains of approximately 83.9 percentage points in long-context understanding and agent-related evaluations.<\/p>\n

SKT also applied its proprietary Sparse Gated Attention (SGA) architecture to A.X K2. The architecture selectively references the most relevant information as the model processes long contexts, helping maximize accuracy and operational efficiency in industrial environments.<\/p>\n

A.X K2 demonstrated performance on par with or superior to Qwen 3.5-397B and DeepSeek-V4-Flash \u2014 both released within the past six months\u2014 on math benchmarks (AIME26, Apex) and Korean-language (KMMLU-Pro, CLIcK) benchmarks.\u201d<\/p>\n

Supporting AI Transformation Across Manufacturing, Public Sector and Beyond with A.X K2<\/h2>\n

SKT plans to expand the application of A.X K2 across industry and everyday life. The company conducted four rounds of post-training to verify the model\u2019s response quality, stability, and operational efficiency, and is identifying use cases across various sectors, including manufacturing, defense, and bio.<\/p>\n

\u2022 Manufacturing AI:<\/strong> SKT is embedding operational know-how from industrial sites into manufacturing-specialized AI agents powered by A.X K2. In the second half of the year, the company plans to deploy demo versions at steelmaker KG Steel\u2019s cold-rolling line in Dangjin, followed by field trials. Similar trials will also be conducted across auto parts manufacturer KONEC\u2019s casting and machining processes.\u201d<\/p>\n

\u2022 Defense AI:<\/strong> In collaboration with the Ministry of National Defense under the Sovereign AI Foundation Model project, SKT has developed three quantized models based on A.X K1. Given the defense sector\u2019s need for the highest levels of security and data sovereignty, the models apply quantization technology to reduce memory usage and increase processing speed. The Ministry and SKT plan to deploy the models across units under the Army, Navy, Air Force, and Marine Corps.<\/p>\n

\u2022 Bio AI:<\/strong> SKT\u2019s proprietary foundation model is also being considered for applications in bio. The company is applying AI to the development of targeted therapies for intractable cancers with SK Biopharmaceuticals, recently helping reduce the early-stage drug research period \u2014 typically one to two years using conventional methods \u2014 to five months.<\/p>\n

\u2022 Enterprise AI Agents:<\/strong> For enterprise productivity, A.X K2 powers A. Biz, enabling users to generate newsletters, gather and analyze information, and create reports. SKT also provides a lightweight version of A.X K2 for SK hynix\u2019s internal tools to support document drafting and information management tasks.<\/p>\n

SKT is also supporting the domestic AI ecosystem by providing A.X K1 in API form to participants in programs led by the Ministry of Science and ICT and to three startups \u2014 namely Purselly (Marketing), REmited (Commerce), and Devall Company (Finance) \u2014 selected for the Ministry of SMEs and Startups-led open innovation program.<\/p>\n

A lightweight version of A.X K2 has also been integrated into SKT\u2019s AI service, A. (A-DoT), expanding its use into everyday applications. A.X K2 powers the \u201cAI Suggestions\u201d feature in A. Phone, which recommends schedules and to-do items based on users\u2019 call conversations. The model is also used in A. Note\u2019s content summarization features.<\/p>\n

Unveiling Vision and Audio Models to Scale Manufacturing AI and Service Adoption<\/h2>\n

Centered on A.X K2, SKT has built a derivative lineup \u2014 including a vision-language (VL) model based on a vision encoder and a speech model based on an audio encoder\u00b7decoder \u2014 to meet diverse industry needs.<\/p>\n

A.X K2 VL Light-Preview and A.X K2 Audio Language Model (ALM) were developed by SKT, while A.X K2 Raon-Speech was developed in collaboration with Krafton, building on SKT\u2019s A.X K2 Light model.<\/p>\n

\u2022 A.X K2 VL Light-Preview (21B):<\/strong> Combines image and text understanding to support enterprise workflows involving reports, manuals and field imagery. Potential applications include analyzing image-based documents such as receipts and academic papers, identifying risks in legal documents, and interpreting industrial piping.<\/p>\n

\u2022 A.X K2 ALM (22B):<\/strong><\/a> <\/span>Designed to recognize and analyze audio from customer consultations, meetings and field environments. SKT is evaluating its use in contact center voice calls.<\/p>\n

\u2022 A.X K2 Raon-Speech (21B):<\/strong><\/a><\/span> Planned for use in Krafton games to enhance voice-based interactions between players and AI.<\/p>\n

Going forward, SKT plans to scale the successor to A.X K2 to the trillion-parameter range to secure global competitiveness. The company has been expanding its parameter scale to match that of leading global players, and introduced 500B-plus large-scale models in the first phase of the Sovereign AI Foundation Model Project.<\/p>\n

\u201cWe have advanced a range of capabilities in parallel to enable broad use across everyday life, offices and manufacturing sites,\u201d said Kim Tae-yoon, Head of Foundation Model Office at SKT. \u201cWe will continue expanding the adoption of A.X K2 across industries to help drive Korea\u2019s AI competitiveness.\u201d<\/p>\n

\u203b Attachment. Hugging Face Links for A.X K2 and Derivative Models<\/strong>
\n\u2013 A.X K2: https:\/\/huggingface.co\/skt\/A.X-K2<\/a><\/span>
\n\u2013 A.X K2 ALM: https:\/\/huggingface.co\/skt\/A.X-K2-ALM<\/a><\/span>
\n\u2013 A.X VE (Vision Encoder): https:\/\/huggingface.co\/skt\/A.X-VE<\/a><\/span>
\n\u2013 A.X K2 Raon-Speech: https:\/\/huggingface.co\/KRAFTON\/A.X-K2-Raon-Speech-21B-A3B<\/span><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

\u00b7\u00a0 Unveils 688-billion-parameter model A.X K2 alongside three lightweight models for advanced multimodal capabilities \u00b7\u00a0 Expands AI across manufacturing, defense and bio while supporting the wider AI ecosystem, including startups \u00b7\u00a0 Develops vision and audio derivative models in parallel to<\/p>\n","protected":false},"author":2,"featured_media":3208,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[4],"tags":[558,24],"class_list":["post-3204","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-press-release","tag-a-x-k2","tag-ai"],"acf":[],"_links":{"self":[{"href":"https:\/\/news.sktelecom.com\/en\/wp-json\/wp\/v2\/posts\/3204","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/news.sktelecom.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/news.sktelecom.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/news.sktelecom.com\/en\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/news.sktelecom.com\/en\/wp-json\/wp\/v2\/comments?post=3204"}],"version-history":[{"count":2,"href":"https:\/\/news.sktelecom.com\/en\/wp-json\/wp\/v2\/posts\/3204\/revisions"}],"predecessor-version":[{"id":3210,"href":"https:\/\/news.sktelecom.com\/en\/wp-json\/wp\/v2\/posts\/3204\/revisions\/3210"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/news.sktelecom.com\/en\/wp-json\/wp\/v2\/media\/3208"}],"wp:attachment":[{"href":"https:\/\/news.sktelecom.com\/en\/wp-json\/wp\/v2\/media?parent=3204"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/news.sktelecom.com\/en\/wp-json\/wp\/v2\/categories?post=3204"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/news.sktelecom.com\/en\/wp-json\/wp\/v2\/tags?post=3204"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}