Specifications
Platform
PINE64 StarPro64 RISC-V single-board computer
SoC
ESWIN EIC7700X
NPU
20 TOPS@INT8
Preinstalled Model
7B DeepSeek / Qwen LLM (Docker container)
Onboard Storage
64 GB eMMC module (pre-loaded)
LLM Runtime
Docker-based inference stack
CPU
RISC-V quad-core, up to 1.8 GHz
Memory
LPDDR5 (StarPro64 configuration)
Networking
Dual Ethernet ports
PCIe
4× PCIe lanes
USB
USB 2.0 and USB 3.0 ports
Video Output
Digital video output (HDMI-class)
GPIO
Full GPIO header
Board Format
Star64 model-A compatible
OS
Linux (RISC-V)
Overview
The PINE64 ALPHA-One is a Generative 7B LLM AI Agent based on the StarPro64 RISC-V single-board computer. It pairs a 20 TOPS INT8 NPU with a preinstalled 64 GB eMMC module that ships with a 7B DeepSeek / Qwen large language model build in Docker form.
It is a ready-to-run reference for on-device generative AI on open RISC-V silicon — no x86 host and no discrete GPU required.
Key Benefits
On-device 7B LLM inference on open RISC-V with a 20 TOPS NPU and a preloaded Docker LLM stack — a turnkey starting point for private, local generative AI without cloud dependency.
Applications
Private on-device LLM inference, edge AI agents, sovereign/offline AI deployments, RISC-V AI development and education, and data-sensitive environments.
Request a Quote — PINE64 ALPHA-ONE — RISC-V GENERATIVE AI AGENT (7B LLM)
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →