← All hardware
Run local LLMs on a Raspberry Pi 5 for $130.
Pros
- On-device generative AI on a sub-$200 platform
- 8GB dedicated RAM unusual at this price point
- Leverages the massive Raspberry Pi ecosystem
Cons
- Limited to smaller (~1-1.5B) models in practice
- Requires a Raspberry Pi 5 host
- Newer product with maturing model support
✓ Where it shines / best for
- Privacy-first local LLM/VLM projects
- Makers experimenting with generative AI at the edge
- Education and offline AI assistants on the Pi 5
✕ Not the best fit for
- Large 7B+ model inference at high speed
- Non-Pi platforms
- High-throughput multi-camera robotics
Features
- ✓ Hailo-10H neural accelerator with 40 INT4 TOPS
- ✓ 8GB dedicated on-board RAM for generative models
- ✓ Computer-vision performance equivalent to the 26 TOPS AI HAT+
- ✓ Runs DeepSeek-R1-Distill, Llama 3.2 and Qwen models locally
- ✓ Backward compatible with original AI HAT+ software
- ✓ Connects via Pi 5 PCIe interface
- ✓ Supports LLMs and vision-language models on-device
- ✓ No cloud dependency for inference
Pricing
| Plan | Price | Billing | Notes |
|---|---|---|---|
| AI HAT+ 2 (Hailo-10H, 40 TOPS INT4, 8GB) | $130 | one-time | Single SKU; available from Raspberry Pi store and approved resellers |
Pricing verified from the official source. Prices change often — confirm on the vendor's site before buying.
Specifications
| power | Low (PCIe-powered HAT) |
| memory | 8GB dedicated on-board RAM |
| interface | PCIe to Raspberry Pi 5 |
| architecture | Hailo-10H neural accelerator on Pi 5 HAT |
| ai_performance | 40 INT4 TOPS |
Sponsored
A full review is being generated for this product and will appear here shortly.