🧩 The LLM630 Compute Kit is a high-performance AI inference development platform designed for edge computing and smart interactive applications. Powered by the AX630C SoC, it integrates a 3.2 TOPs@INT8 NPU for efficient execution of computer vision (CV) and large language model (LLM) tasks.
🌐 Equipped with the JL2101-N040C Gigabit Ethernet chip and ESP32-C6 Wi-Fi 6 module, it supports high-speed data transmission and seamless Wi-Fi–Ethernet bridging. The SMA antenna interface enhances wireless stability and range in complex environments.
💾 The kit includes 4GB LPDDR4 RAM (2GB user + 2GB acceleration) and 32GB eMMC storage, enabling parallel model loading and smooth multi-task inference.
🧠 The baseboard expands functionality with a BMI270 6-axis sensor for motion detection, NS4150B Class D amplifier, and microphone/speaker interfaces for full-duplex audio interaction. It supports MIPI DSI/CSI interfaces for display and camera modules, dual Grove ports, and a Gigabit Ethernet port.
🔋 Integrated charging and battery detection chips, plus a reserved battery socket, ensure long-term operation. The MicroSD slot supports storage expansion and future model updates. Dual Type-C ports offer high-speed data and OTG functionality.
🚀 Compatible with the StackFlow framework, developers can launch edge AI tasks with minimal code. Supported models include Yolo11, DepthAnything, InternVL2.5-1B, Qwen2.5, Llama3.2, and Whisper Melotts, with hot-update support for future LLMs and multimodal AI.
🔧 Ideal for security monitoring, smart retail, agriculture, home automation, robotics, and AI education.
The StackFlow framework is an open-source AI framework developed for the Module LLM and LLM630 Compute Kit. It integrates AI algorithms such as speech, vision, and large language models.
🔧 Features
- AX630C SoC with 3.2 TOPs @ INT8
- Gigabit Ethernet + ESP32-C6 Wi-Fi 6
- Display (DSI) and Camera (CSI) interfaces
- Charging and battery detection chips
- Microphone and speaker interfaces
- MicroSD slot for storage expansion
- 32GB eMMC + 4GB LPDDR4 RAM
- Dual Type-C ports with OTG
- SMA antenna interface
- User buttons for power and mode
- Dual Grove ports (I2C + UART)
- BMI270 6-axis sensor
- StackFlow development platform
📦 Package Includes
- 1 × LLM630 Compute Kit
- 1 × SMA 2.4G Wi-Fi Antenna
- 1 × Heatsink
- 4 × EVA Foam Foot Pads
🚀 Applications
- Smart Home
- Security Monitoring
- Smart Retail
- Smart Agriculture
- Interactive Robotics
- Education & Learning Platform
📊 Specifications
| Specification | Parameter |
|---|---|
| Processor | AX630C @ Dual Cortex A53 1.2GHz 12.8 TOPs @ INT4, 3.2 TOPs @ INT8 |
| NPU | 3.2 TOPs @ INT8 |
| RAM | 4GB LPDDR4 (2GB system + 2GB acceleration) |
| eMMC | 32GB eMMC 5.1 |
| Wired Network | JL2101B-N040C @ 1GbE |
| Wireless Network | ESP32-C6 @ Wi-Fi 6 2.4GHz |
| USB-UART | CH9102F |
| USB-OTG | USB 2.0 Host/Device |
| Antenna Port | SMA inner hole |
| Audio Interface | MIC & SPK Header 5P @ 1.25mm |
| Display Interface | MIPI DSI 2Lane, 1080p@30fps |
| Camera Interface | MIPI CSI 4Lane, 4K@30fps |
| Battery Management | AW32001ECSR + BQ27220YZFR |
| Battery Interface | 1.25mm spec terminal |
| Battery Spec | 3.7V Li-ion/Li-polymer |
| USB Ports | 2 × Type-C (data + OTG) |
| USB Power Input | 5V @ 2A |
| Grove Interface | PortA (I2C), PortC (UART) |
| Storage Expansion | MicroSD card slot |
| External Interface | FUNC Header 8P @ 1.25mm |
| Buttons | 2 × user/power/reset |
| Sensor | BMI270 6-axis |
| Operating Temp | 0°C – 40°C |
| Product Size | 90.3 × 31.6 × 12 mm |
| Package Size | 133 × 83 × 17 mm |
| Product Weight | 22.9 g |
| Package Weight | 60.0 g |