AIBOX-1688 Overview Specifications Wiki Technical Case Download
Buy NowDataSheet
Buy NowDataSheet
AIBOX-186 Overview Specifications Wiki Technical Case Download
Buy NowDataSheet
Buy NowDataSheet

Putting Large Models
into a Small Box

Low power consumption AI box
——AIBOX-186

7.2 TOPS of computing power!
Support the private deployment of mainstream large models.
Bring private AI capability to meet individual AI deployment needs.

Comprehensive private AI deployment

7.2 TOPS AI processor CV186AH

CV186AH, the SOPHON AI processor, features an Hexa-core ARM Cortex-A53 with up to 1.6 GHz of frequency. Equipped with a neural network acceleration
engine TPU, it delivers peak performance of 7.2T@INT8 and 12T@INT4, 1.5T@FP16/BF16,. With support for mainstream programming frameworks,
this processor can be widely used in AI inference, computer vision, and more.

Multi-channel video AI performance

The AI box supports up to 16 channels of H.265 / H.264 1080P video decoding, 10 channels of H.265 / H.264 1080P video encoding,
and 16 channels of 1080P HD video processing (decoding + AI analysis). This meets the needs of various AI applications
such as face detection on video streaming, license plate recognition, and smart cities.

Strong network communication capability

With dual 1000Mbps Ethernet, the AI box ensures high-speed and stable network communication,
meeting the needs of various application scenarios.

All-aluminum alloy enclosure for heat dissipation

The AI box features an industrial-grade all-metal enclosure with an aluminum alloy structure for thermal conduction. The side of the top cover
features a grille design for external airflow and efficient heat dissipation, ensuring computing performance and stability
even under high-temperature operating conditions.

Abundant resources

We offer SDKs, tutorials, technical documentation, and development tools to streamline and improve the development process.

A wide range of applications

AIBOX-186 is widely used in intelligent surveillance, AI education, services based on computing power, edge computing,
private deployment of large models, and data security and privacy protection.

Intelligent
surveillance
AI education
Computing power
services
Edge computing
private deployment
of large models
Data security

Specifications

AIBOX-1688 AIBOX-186

Basic Specificat

SOC

SOPHON BM1688

SOPHON CV186AH

CPU

Octa-core ARM Cortex-A53 @ 1.6GHz

Hexa-core ARM Cortex-A53 @ 1.6GHz

TPU

Built-in SOPHGO neural network acceleration engine TPU, 32T@INT4 peak computing power, 16T@INT8 peak computing power, 4T@FP16/BF16computing power, 0.5T@FP32 computing power

7.2T@INT8, 12T@INT4, and 1.5T@FP16/BF16 computing power

Decoding/ Encoding

Video decoding: H.264 / H.265 decoding (Max performance: 1920 * 1080@480FPS or 3840 * 2160 @120FPS) Video encoding: H.264 / H.265 encoding (Max performance: 1920 * 1080@300FPS or 3840 * 2160 @75 FPS) Image codec: JPEG/MJPEG Baseline codec (JPEG codec with a maximum resolution of 1080P@480 FPS)

RAM

8GB LPDDR4 (4GB/8GB/16GB optional)

4GB LPDDR4 (4GB/8GB/16GB optional)

Storage

32GB eMMC (32GB/64GB/128GB/256GB optional)

Storage Expansion

1*M.2 (Expandable PCIe NVMe SSD(default support)/ SATA SSD(supported after software update), supports 2242/2260/2280) (inside the device),1*TF Card

Power

DC 12V/3A(DC 5.5*2.1mm)

Power consumption

Normal: 7.2W(12V/600mA), Max: 14.4W(12V/1200mA)

Normal: 6W(12V/500mA), Max: 10.8W(12V/900mA)

OS

Linux

Software Support

・ The private deployment of ultra-large-scale parameter models under the Transformer architecture, 　　including large language models such as Gemma-2B, LlaMa2-7B, ChatGLM3-6B, Qwen1.5-1.8B. ・ Traditional network architectures such as CNN, RNN, and LSTM; a variety of deep learning frameworks, 　　including TensorFlow, PyTorch, MXNet, PaddlePaddle, and ONNX, as well as custom operator development ・ Docker container management technology

・The private deployment of ultra-large-scale parameter models 　　under the Transformer architecture, including large language 　　models such as Gemma-2B, LlaMa2-7B, ChatGLM3-6B, 　　Qwen1.5-1.8B. ・ A variety of deep learning frameworks, including 　　TensorFlow, PyTorch, TensorRT, TFLite, PaddlePaddle, 　　Caffe and ONNX, as well as custom operator development ・ Docker container management technology

Size

93.4mm × 93.4mm × 50 mm

Weight

≈ 500g

Environment

Operating Temperature: -20℃～60℃, Storage Temperature: -20℃～70℃, Storage Humidity: 10%～90%RH (non-condensing)

Interface Specif

Ethernet

2*1000Mbps Ethernet

Video output

1*HDMI2.0（4K@60fps)

USB

2*USB3.0 (Current Limit: 1A)

Other

1*Type-C (USB 2.0 device mode only), 1*Console (Debug serial), 1*Power button, 1*Recovery