Anonymous
Mar 31, 2026
Rating: 5/5
Industrial grade performance
We evaluated this for an industrial inspection system. The thermal management is good even under sustained load. The GPIO headers are well labeled and easy to interface with sensors.
Description
Reviews
| Brand | - |
| Category | Robotics / STEM Teaching Tools / Development, Learning, Evaluation & Industrial Control Boards |
| Origin | - |
| Processor | Qualcomm QCS6490 Octa-core |
| AI Performance | 12 TOPS (INT8) |
| Memory Type | LPDDR5 |
| Memory Capacity | 8GB |
| Storage | 32GB eMMC + microSD slot |
| Connectivity | Wi-Fi 6, Bluetooth 5.2, Gigabit Ethernet |
| Display Output | HDMI 2.0, MIPI DSI |
| Camera Interface | 2x MIPI CSI |
| Operating System Support | Android 12, Ubuntu Linux |
| Power Supply | USB Type-C (5V/3A) |
Anonymous
Mar 31, 2026
Rating: 5/5
Industrial grade performance
We evaluated this for an industrial inspection system. The thermal management is good even under sustained load. The GPIO headers are well labeled and easy to interface with sensors.
Anonymous
Mar 18, 2026
Rating: 5/5
Best value for computer vision
Compared to other boards in this price range, the NPU performance is unmatched. We deployed our proprietary image classification algorithm and it runs smoothly at 60fps.
Anonymous
Mar 17, 2026
Rating: 4/5
Steep learning curve but worth it
As a student, this was my first dive into Qualcomm architecture. It's more complex than Raspberry Pi but the performance gain is massive. Great for learning advanced embedded systems.
Anonymous
Mar 17, 2026
Rating: 5/5
Perfect for autonomous robot prototypes
Used this for a small delivery robot project. The dual camera interfaces allowed for stereo vision setup easily. Wi-Fi 6 connectivity ensures low latency control. Highly recommended for robotics.
Anonymous
Mar 07, 2026
Rating: 5/5
Excellent board for edge AI projects
The 12 TOPS computing power is no joke. I successfully ran multiple object detection models simultaneously without lag. The LPDDR5 memory makes a huge difference in data throughput compared to older boards.
Anonymous
Mar 06, 2026
Rating: 4/5
Great hardware, documentation needs work
Hardware is solid and the Qualcomm chip performs well. However, getting the custom Linux image set up took some time as the wiki pages are a bit sparse. Once running, it's very stable.
Q: Does it come with a heatsink or cooling fan?
A: The standard kit includes a passive aluminum heatsink. For heavy AI workloads, we recommend purchasing the active cooling fan accessory separately.
Q: Is the source code for the BSP (Board Support Package) open source?
A: Yes, Radxa provides the kernel source and device tree files on our GitHub repository. Some proprietary Qualcomm binaries are included as required for full functionality.
Q: Does this board support Android 13 or only Android 12?
A: Currently, the official image supports Android 12. An update to Android 13 is in beta testing and expected to be released in the next quarter.
Q: What is the maximum resolution supported by the MIPI CSI cameras?
A: The board supports up to 4K resolution at 30fps or 1080p at 60fps per camera channel, depending on the specific sensor module used.
Q: Can I power this board via the GPIO pins instead of USB-C?
A: While there are 5V pins on the header, we strongly recommend using the USB Type-C port for primary power to ensure stable voltage regulation and prevent damage to the board.