APMIC announced today (April 4) that the lightweight 29B parameter language model "Formosa-3" developed with the Twinkle AI community is now deeply integrated with the Model Context Protocol (MCP).
"Formosa-1" is Taiwan's first 3B parameter Traditional Chinese language model that can run on mobile phones. Now, through the MCP protocol, Formosa-1 can connect to internal and external enterprise systems, enabling natural language-driven information query, process operation and decision support, creating a new intelligent human-computer interaction interface, and also supports connection with external APIs, ERP systems and IoT devices.
Combined with the fine-tuning and distillation solutions provided by the APMIC S1 model foundry service, enterprises can build customized edge models that run smoothly and respond instantly, whether deployed on mid-range or low-end GPUs or embedded platforms. This significantly reduces the technical and cost barriers to implementing AI solutions, meeting the demands of industrial equipment manufacturers, technology companies, mobile device brands, and robotics and human-machine interface (HMI) applications for edge deployment, low power consumption, and fast response.
Natural language moves from "communication" to "control"
Unlike traditional chat-centric language models, Formosa-1 integrates the MCP protocol, enabling deep integration with internal enterprise systems. Driven by language, the model not only responds to queries but also serves as an intelligent command interface for triggering business processes, accessing data, and supporting decision-making, further expanding the practical application of natural language.
This solution has demonstrated practical benefits in a variety of scenarios. For example, in industrial automation, engineers can use voice to query CNC error codes or activate specific module equipment. Retailers can use voice to operate POS systems, quickly accessing sales data and inventory information. In logistics control, language models can instantly query inventory status and assign delivery tasks to specific drivers, significantly improving operational efficiency. Using the MCP protocol, Formosa-1 can flexibly connect to external APIs such as ERP systems, weather data, exchange rate information, and machine monitoring. This allows users to trigger complex operational processes through natural language, replacing traditional interfaces or manual input to enhance interactive intuitiveness.
APMIC "Formosa-1" can now achieve deep integration with internal enterprise systems through the MCP protocol, comprehensively expanding the application boundaries of language models.
APMIC Founder and CEO Wu Bohan said, "Promoting the low-cost deployment of AI models across diverse devices will be a key path to achieving AI ubiquity. APMIC's independently developed 3B model, combined with the MCP protocol, represents a significant step in moving human-computer interaction from a 'cloud dream' to an 'embedded reality.' APMIC has long been committed to developing localized language models and enterprise-private AI deployment solutions. The 'Formosa-1' Traditional Chinese model demonstrates APMIC's strength in model refinement and field deployment. The integration of the 3B model and the MCP protocol at this stage demonstrates APMIC's continued advancement of AI technology into more fields through practical applications, enabling more enterprises, scenarios, and devices to truly realize the ability for language commands to drive real-time business actions."
The "Formosa-1" Traditional Chinese model is linked to the exchange rate information MCP protocol, allowing you to understand real-time exchange rates through commands.
Lightweight deployment architecture enables instant operation of language models on the ground
Formosa-1 utilizes a 3B-scale design, offering high deployment flexibility. It can run smoothly on mobile phones, IoT devices, robots, human-machine interfaces (HMIs), and various edge devices, eliminating the need for cloud computing resources or high-end GPUs, significantly lowering the barrier to entry. This feature allows businesses to flexibly deploy language models on mobile devices and remote locations, expanding the application of intelligent technology in diverse fields such as manufacturing, retail, and logistics.
While not specifically designed for deep logical reasoning, Formosa-1 demonstrates exceptional performance in processing structured queries, automating task processes, and integrating multi-modal data. To help businesses customize models for specific industry needs, APMIC also launched the Embedded Business Model Fine-Tuning solution. This solution, based on internal knowledge bases and operational processes, can create tailored models with industry-specific language and command understanding, further enhancing the accuracy and value of models in practical applications.
