
Jetson Orin Nano Core Board (8GB) J127*1
A compact generative AI supercomputer developer kit designed for commercial AI developers, enthusiasts, and students.
Significant performance upgrade over previous generation with 1.7x generative AI inference performance improvement.
Enhanced compute power with 67 INT8 TOPS for 8GB model and 34 INT8 TOPS for 4GB model, representing a 70% increase.
Memory bandwidth increased by 50% to 102 GB/s, supporting large-scale AI inference workloads.
Ideal for building retrieval-augmented LLM chatbots, visual AI agents, and AI-powered robotics.
| Item | Detail |
|---|---|
| Orin Nano 8GB Super | 8GB LPDDR5, 128-bit, 102 GB/s bandwidth |
| Orin Nano 8GB | 8GB LPDDR5, 128-bit, 68 GB/s bandwidth |

TYPEC interface 168P chip soldered + cable
CN¥22.68
| Orin Nano 4GB Super | 4GB LPDDR5, 128-bit, 51 GB/s bandwidth |
| Orin Nano 4GB | 4GB LPDDR5, 128-bit, 34 GB/s bandwidth |
Large Language Model (LLM) performance improvements ranging from 1.28x to 1.63x across popular models including Llama 3.1/3.2, Qwen 2.5, Gemma 2, Phi 3.5, and SmolLM2.
Enhanced GPU and CPU utilization with improved clock speeds and thermal management.
Supports AI large model voice module with onboard high-performance dual microphone array enabling far-field voice pickup, echo cancellation, and environmental noise reduction.
Enables natural and smooth human-machine voice interaction through integration with large language and speech models.
Pre-installed Ubuntu 20.04-based Jetson Linux environment with all necessary AI libraries and tools for immediate development.
Precise motherboard and mounting hole dimensions provided for easy expansion and peripheral development.
Compact form factor suitable for handheld use.
AI-powered voice interaction terminals
Visual AI agents and object detection
Robotics and autonomous systems
Generative AI model development and deployment
All software and hardware components are optimized for high concurrency and high compute workloads.
Strict copyright enforcement on development guide content to protect intellectual property.
© 2024 RB