null

CM4 Edge AI Compute Module - Radxa

Radxa

No reviews yet Write a Review
SKU:
498-RM126-D4E32J0R35
MPN:
RM126-D4E32J0R35
Radxa CM4 Edge AI Compute Module with Rockchip RK3576, octa-core CPU, Mali-G52 GPU, 6 TOPS NPU, LPDDR4x up to 16GB, and eMMC up to 256GB. Ideal for industrial AI, edge computing, vision, and multimedia. View full description
  • 100% Original Products
  • GST Invoice
  • Earn 3% Cashback
₹7,727.82 inc. GST
₹6,549.00 ex. GST
4 people looking at this.

Technical Specifications

Brand:
Radxa
SoC:
Rockchip RK3576 or RK3576J
CPU:
Quad‑core Cortex‑A72 (1MB L2) and Quad‑core Cortex‑A53 (512KB L2) ARMv8‑A, TrustZone, Cryptography Extensions
GPU:
ARM Mali G52 MC3
NPU:
6 TOPS (INT8), supports INT4/INT8/INT16/FP16/BF16/TF32 compatible with TensorFlow, PyTorch, ONNX, etc.
Memory:
LPDDR4/LPDDR4X (up to 4266MT/s) dual‑channel 16bit, up to 16GB total address space
Codec:
H.264 multivideo decoder up to 4K 60fps,H.265 multivideo encoder up to 4K 60fps
Storage:
Onboard eMMC Up to 256GB,
Storage Optional onboard SPI Flash:
1x UFS 2.0 ,1x SDMMC
Ethernet:
1x Onboard Gigabit Ethernet PHY supporting IEEE 1588
Video Input:
1x 4‑lane MIPI CSI,1x 4‑lane MIPI CSI or 2x 2‑lane MIPI CSI
Video Output:
1x HDMI TX up to 4K 120Hz,1x 4‑lane eDP Interface up to 4K60fps,1x DP Interface up to 4K 120fps,1x 4‑lane MIPI DSI for MIPI LCD,Able to provide 2x display output
USB:
1x USB 3.2 Gen1x1 OTG / HOST,1x USB 3.2 Gen2x1
PCIe:
2x PCIe Gen2 1‑lane
Low Speed Buses:
SPI, I2C, I3C, UART, CAN
Operation Temperature:
0C to 60C(Commercial Grade) ‑40C to 85C(Industrial Grade)

Warranty Information

All the products supplied by Evelta are genuine and original. We offer 14 days replacement warranty in case of manufacturing defects. For more details, please visit our cancellation and returns page.

Description

Resources:

Datasheet

CM4 Edge AI Compute Module – Compact, Powerful, and Ready for Deployment: Accelerate your next-generation embedded solutions with advanced AI computing

The Radxa CM4 Edge AI Compute Module is a high-performance system-on-module engineered for industrial control, edge computing, computer vision, and multimedia applications. Powered by the Rockchip RK3576 or RK3576J SoC, this compact 55 mm x 40 mm module combines an octa-core CPU (4x Cortex-A72 + 4x Cortex-A53), ARM Mali-G52 GPU, and a 6 TOPS NPU for advanced AI inference. It supports modern AI frameworks like TensorFlow, PyTorch, ONNX, and Caffe, making it ideal for developers building AI-driven solutions with industrial-grade reliability.

With flexible connectivity options, onboard storage up to 256GB, and support for LPDDR4x RAM up to 16GB, the CM4 ensures seamless integration into space-constrained custom embedded designs. Multimedia capabilities include AV1/VP9/H.265 decoding up to 8K@30fps, H.264 up to 4K@60fps, and dual-display output support via HDMI, eDP, DP, or MIPI interfaces. Designed for long-term deployment, the module is supported across Debian, Linux, Buildroot, and Android, making it a versatile platform for developers and enterprises alike.

Key Features

  • High-Performance SoC: Rockchip RK3576(J) with Quad-Core Cortex-A72 + Quad-Core Cortex-A53, up to 2.2GHz

  • Powerful NPU: 6 TOPS@INT8 with support for INT4/INT8/INT16/FP16/TF32 and AI frameworks including TensorFlow, ONNX, PyTorch, and Caffe

  • Advanced GPU: Mali-G52 MC3 with OpenGL ES 1.1/2.0/3.2, Vulkan 1.2, and OpenCL 2.1 support

  • Flexible Memory Options: LPDDR4X RAM up to 16GB and eMMC storage up to 256GB with support for UFS 2.0, SPI Flash, and SDMMC

  • Comprehensive Multimedia Support: Hardware decoding up to 8K@30fps, dual 4K display output, and encoding up to 4K@60fps

  • Robust Connectivity: Onboard WiFi 6, Bluetooth 5.4, Gigabit Ethernet PHY, PCIe, SATA, USB 3.2 Gen2x1, GPIO, I2C, SPI, UART, PWM, and more

  • Industrial Reliability: Operating range from -40°C to 85°C with three 100-pin B2B connectors for secure integration

  • Wide OS Support: Debian, Linux, Buildroot, and Android for flexible software development