Hon Hai unveils new traditional Chinese language AI model

Taipei, March 10 (CNA) Hon Hai Precision Industry Co. said Monday that its research institute has launched its first traditional Chinese version of an advanced artificial intelligence Large Language model (LLM), with technology assistance from U.S.-based AI chip designer Nvidia Corp.

Hon Hai Precision Industry Co., which assembles iPhones and also rolls out AI servers in close cooperation with Nvidia, said the LLM, coded FoxBrain, is expected to strengthen its data analysis capabilities for future smart manufacturing, and electric vehicle and smart city development.

An LLM is a type of machine learning model designed for natural language processing tasks such as language generation, and it plays an important role in the development of AI servers.

Nvidia provided assistance and consultation through its "Taipei-1" supercomputer, allowing Hon Hai Research Institute to use the Nvidia NeMo AI service platform to complete the training of the FoxBrain LLM, Hon Hai Precision said in statement.

During the training process of the FoxBrain platform, the AI research team used 120 Nvidia H100 graphics processing units (CUPs) and Nvidia Quantum-2, the seventh generation of the Nvidia InfiniBand, which gives AI developers and scientific researchers the fastest networking performance, Hon Hai said.

The training was completed in only four weeks at a low cost but with high efficiency, according to Hon Hai, the world's largest contract electronics manufacturer.

The FoxBrain model shows understanding and reasoning capabilities, performs well in mathematics and logical reasoning tests, and enhances the language style of Taiwanese users, the Taiwanese company said.

The model covers a wide range of areas such as data analysis, decision making aid, writing assistance, mathematics, code generation, and problem solving through its inferential capability, according to the statement.

In the future, FoxBrain will be made open source, providing AI computing codes to other enterprises and researchers, thus allowing them to use LLM technologies to promote their own products, Hon Hai said.

The FoxBrain platform has made significant progress in math tests, compared with the top ranked Chinese language model Taiwan Llama, and it surpasses Meta's current models at the same level in terms of mathematical reasoning ability, according to Hon Hai.

While FoxBrain is still a little behind the China-based AI startup DeepSeek, its performance is already close to the best in the world, Hon Hai said.

The FoxBrain platform is expected to drive the growth of smart devices and help advance applications in manufacturing, supply chain management and smart decision making, based on AI development, according to the company.

The Taiwanese manufacturing giant is scheduled to highlight FoxBrain at Nvidia's GPU technology conference in San Jose, California, which will kick off March 17. Hon Hai will present the achievements in a keynote speech titled "From Open Source to Frontier AI: Build, Customize, and Extend Foundation Models."

In recent years, Hon Hai has intensified its efforts to diversify its product portfolio through its "3 plus 3" strategy in a bid to transition from solely contract manufacturing into hardware and software fields.

The initiative covers three emerging industries -- EVs, robots and digital health care -- which the company said it is developing via AI, semiconductor and communications technologies, with EVs as the core business.