Integrated GPU — The Sweet Spot
A compact powerhouse with AMD's integrated graphics delivering real GPU-accelerated inference — 3–5× faster than CPU-only — plus the ability to run image generation and 24/7 AI agents from a single, silent box.
~15–25
tokens / second
Integrated GPU inference
30s
per image (CPU)
Stable Diffusion · CPU mode
24/7
Agent uptime
Low power · always on
Price
~$1,000
Perfect for: home workers · small business · students · anyone who wants speed and image generation without a desktop tower
What you get
- AMD integrated GPU — real hardware acceleration for LLMs
- CPU-based image generation (Stable Diffusion, FLUX)
- 24/7 AI agent operation — always listening, always working
- Runs 7B–14B models comfortably at chat speed
- Compact, silent, low power (~25–65 W)
- Pre-configured with Ollama, Open WebUI, RAG pipeline
Technical Specs
| CPU | AMD Ryzen 5 (6-core, with integrated Radeon GPU) |
|---|---|
| Integrated GPU | AMD Radeon 760M or 780M iGPU |
| RAM | 16–32 GB DDR5 (shared with GPU) |
| Storage | 256 GB SSD (expandable) |
| LLM inference | ~15–25 tok/s (iGPU accelerated) |
| Image generation | ~30–60 s per image (CPU) |
| Agent capability | 24/7 — low power draw |
| Power | 25–65 W typical |
| OS | Linux (Ubuntu pre-configured) or Windows 11 |