Specifications
SoC
Allwinner A733
CPU
Octa-core: 2× Cortex-A76 + 6× Cortex-A55, up to 2.0 GHz
GPU
Imagination BXM-4-64
NPU
3 TOPS@INT8; INT8 / INT16 / FP16 / BF16 mixed precision
AI Frameworks
TensorFlow, PyTorch, ONNX
MCU
RISC-V E902 co-processor @ 200 MHz
PMIC
AXP318
Memory
Up to 16 GB LPDDR5
Storage
eMMC module 16/32/64/128 GB, SPI flash 128/256 Mb, M.2 M-Key NVMe SSD, UFS module, TF card up to 128 GB
Wireless
Wi-Fi 6 + Bluetooth 5.4 (BLE)
Ethernet
Gigabit Ethernet (YT8531CA PHY) with PoE support
Display
1× HDMI 2.0 up to 4K@60Hz + 1× 4-lane MIPI DSI
Camera
1× 2-lane + 1× 4-lane MIPI CSI
USB
1× USB 3.0 + 3× USB 2.0
Expansion
40-pin header (GPIO, UART, I2C, SPI, PWM)
Power
USB Type-C 5V / 3A
Dimensions
89 × 56 mm, 58 g
Overview
The OrangePi 4 Pro is a compact edge-computing board built on the Allwinner A733 SoC. It combines an octa-core CPU (2× Cortex-A76 + 6× Cortex-A55 up to 2.0 GHz) with a 3 TOPS@INT8 NPU supporting INT8/INT16/FP16/BF16 mixed precision and mainstream frameworks, plus an Imagination BXM-4-64 GPU and a RISC-V E902 co-processor for real-time control.
With up to 16 GB of LPDDR5, Wi-Fi 6, PoE-capable Gigabit Ethernet, NVMe and UFS storage options and 4K HDMI output, it suits robotics, smart gateways, industrial control and AI inference appliances.
Key Benefits
3 TOPS NPU with up to 16 GB LPDDR5 and Wi-Fi 6 in an 89 × 56 mm board — a high-value A733 edge platform with PoE, NVMe and framework-ready AI tooling.
Applications
Robotics control with dual-camera vision, industrial automation and PLC gateways, edge AI appliances (face recognition, behaviour detection), smart gateway/NAS and education.
Request a Quote — ORANGEPI 4 PRO — ALLWINNER A733 SBC WITH 3 TOPS NPU
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →