AIBOX PRO KIT is a dual-AI matrix edge computer built for on-device large model deployment. It pairs a modular Rockchip host SoC with one or two independent M.2 AI accelerator modules, so compute power and memory can be scaled by swapping modules instead of redesigning the whole system.
Instead of a fixed-specification AI box, the PRO KIT uses a heterogeneous modular architecture — a main controller plus an independent AI coprocessor — supporting LLM, VLM and AI Agent workloads while remaining compatible with traditional CNN and RNN algorithms. It integrates CAN-FD and RS485 industrial interfaces, making it suitable for both general AI and industrial intelligence deployment.
Key features
Dual-core architecture — host SoC module plus an independent AI coprocessor module
2 × M.2 M-KEY accelerator slots — accepts RK1828, HOUMO LQ50, DeepX DX-M1 and other 2260/2280 M.2 modules
Runs LLMs up to 7B smoothly with RK1828 (5GB DRAM, 20 TOPS INT8); up to 35B with LQ50 (48GB LPDDR5, 160 TOPS INT8)
8K video — 8K@30fps H.265/H.264 encoding and 8K@60fps H.265 decoding with RK3588 host
Onboard SATA 3.0, PCIe NVMe, TF card storage expansion and dual Gigabit Ethernet
Industrial I/O — RS485, CAN-FD, opto-isolated DI/DO, 9–36V wide voltage input, −20°C to 60°C operation
Millisecond-level low latency with rich expansion and full development materials
2 × M.2 M-KEY (supports 2260/2280 accelerator cards & other M.2 modules, inside the device) * Users can install any brand M.2 M-KEY accelerators; standard equipped with RK1828 / HOUMO LQ50 modules ○ RK1828: onboard 5GB DRAM, 20 TOPS (INT8), runs LLMs up to 7B smoothly ○ LQ50: 48GB LPDDR5, 160 TOPS (INT8), runs LLMs up to 35B smoothly
Storage expansion
1 × M.2 M-KEY (expandable with PCIe 2.0 x1 NVMe 2280 SSD; unavailable when Core-3576JD4 serves as the host module) 1 × SATA 3.0 (expandable with SATA 3.0 SSD; unavailable when Core-3576JD4 serves as the host module) 1 × TF card
Power
DC 12V/7A (5.5×2.1mm, supports 9V–36V wide voltage input)
Accelerator power
2 × 12V (2 pcs 4Pin-1.25mm wafer sockets)
Power consumption
RK3588 + 1 × RK1828: Normal 9.6W (12V/800mA), Max 28.8W (12V/2400mA), Min (sleep) 0.72W (12V/60mA) RK3588 + 2 × RK1828: Normal 15.6W (12V/1300mA), Max 42W (12V/3500mA), Min (sleep) 0.72W (12V/60mA) RK3588 + 1 × LQ50: Normal 8.4W (12V/700mA), Max 45.6W (12V/3800mA), Min (sleep) 0.84W (12V/70mA) RK3588 + 1 × LQ50 DUO: Normal 12W (12V/1000mA), Max 63.6W (12V/5300mA), Min (sleep) 0.84W (12V/70mA) RK3576 + 1 × RK1828: Normal 8.4W (12V/700mA), Max 28.8W (12V/2400mA), Min (sleep) 1.2W (12V/100mA)
OS
Linux
Dimensions
148.0mm × 106.0mm × 36.4mm
Environment
Operating temperature: −20°C to 60°C; storage temperature: −20°C to 70°C; storage humidity: 10%–90% RH (non-condensing)
Interface specifications
Internet
2 × Gigabit Ethernet (1000Mbps/RJ45) Wi-Fi & Bluetooth: dual-band 2.4GHz/5GHz Wi-Fi and Bluetooth 5.2 (expandable via M.2 E-KEY; unavailable with Core-3576JD4 host module) 4G: expandable via Mini PCIe | 5G: expandable via M.2 B-KEY
What is AIBOX PRO KIT Dual-AI Matrix Edge Computer used for? AIBOX PRO KIT Dual-AI Matrix Edge Computer is a AI Edge Computer from Firefly. It is typically used in embedded computing, edge AI inference, industrial control and rapid prototyping, and can be integrated into custom carrier boards or finished devices.
Does AIBOX PRO KIT Dual-AI Matrix Edge Computer include technical support, datasheets or pre-flashed images? Yes. Ai-Paipai supplies this item as an authorized distributor and provides local engineering support, datasheets, SDKs and pre-flashed system images where applicable, plus OEM / ODM customization for volume projects.
Can I order AIBOX PRO KIT Dual-AI Matrix Edge Computer with OEM / ODM options or CE / FCC / RoHS certification? OEM / ODM customization (specs, enclosure, BSP) is available, and CE / FCC / RoHS certified options can be provided on request for EU and US imports. Contact us for a quote.
How is AIBOX PRO KIT Dual-AI Matrix Edge Computer shipped, and what is the lead time? Orders ship worldwide from Shenzhen. Stock items typically dispatch within a few business days; configurable or made-to-order variants take longer. Exact lead time is shown at checkout or on request.
Specifications
Host module
RK3588 (Core-3588JD4) or RK3576 (Core-3576JD4) — LPDDR4/LPDDR4x up to 32GB + eMMC up to 256GB, 6 TOPS INT8
AI accelerator slots
2 x M.2 M-KEY (2260/2280)
Standard accelerators
RK1828 — 5GB DRAM, 20 TOPS INT8, runs LLMs up to 7B
Upgrade accelerator
HOUMO LQ50 — 48GB LPDDR5, 160 TOPS INT8, runs LLMs up to 35B