FoxBrain (AI model)

History

2025: Product Announcement

Foxconn has launched its own AI model, FoxBrain, to automate manufacturing processes and manage supply chains. The official presentation of the technology took place on March 10, 2025.

According to the company, FoxBrain was created to improve efficiency at Foxconn plants and optimize work in the field of contract electronics production.

𝓲

Фото: reuters.com

FoxBrain took Reuters only four weeks to develop and train, "" reports. 120 GPUs were used to train the model Nvidia H100 and the largest in the Taiwan supercomputer Taipei-1, owned Nvidia and located in Kaohsiung.

The new AI model is based on the Llama 3.1 architecture developed by Meta (recognized as an extremist company in Russia, its activities are prohibited), and is specially optimized for the traditional Chinese dialect used in Taiwan. FoxBrain became Taiwan's first major language model with logical thinking capabilities adapted to local language features.

Yung-Hui Lee, director of the Foxconn artificial intelligence research center, noted that the model "prioritizes optimized learning strategies over the simple use of computing power" to solve problems.

FoxBrain functionality covers several key areas:

• data analysis and decision support;
• Document collaboration
• mathematical calculations;
• logical reasoning and problem solving;
• program code generation.

The Taiwanese electronics manufacturer, known as the main assembler iPhone for, admitted that Apple FoxBrain is slightly inferior in performance to the Chinese model, DeepSeek but stressed that its overall characteristics are close to world standards.

Originally designed for internal use, the model will be used to improve production processes and optimize Foxconn supply chains. In the future, the company plans to open the source code of FoxBrain, which will allow us to establish cooperation with technological partners to expand the fields of application of artificial intelligence.^[1]

Notes