Guide

What is the cheapest always-on AI assistant setup?

Right now, the cheapest verified always-on build is $1.00/mo (typical $0.60–$1.50): Oracle Always Free (ARM) running GPT-OSS 20B. Always-on means a hosted server, not your own computer. Every number below is dated and sourced.

Monthly breakdown

Oracle Always Free (ARM)	$0.00
GPT-OSS 20B	$1.00
Estimated total	$1.00/mo

typical $0.60–$1.50

Why this stack

Host: Oracle Always Free (ARM). The cheapest verified always-on server that meets the runner's memory floor. Always-on means a hosted server, not your own computer.
Model: GPT-OSS 20B. The lowest full-stack cost among verified, compatible models at moderate daily usage.
Runner: OpenClaw. The default server runner.
Security default on. A private tunnel is included at no cost for every server build.

What changes the cost

Compute: where it runs: a cloud host, a rented GPU, or hardware you own.
Model usage: how much it reads and writes, and how often it runs.
Runner: the software that serves the model, or runs the agent loop and its tools.
Capabilities: browser automation, web search, memory, or scheduled work.
Required extras: security, delivery, monitoring, or managed-host fees.
Freshness: whether each price is verified, aging, stale, or disputed.

Data freshness

Oracle Always Free (ARM):community · 37dGPT-OSS 20B:vendor · 9d

Next-cheapest models on the same host

Granite 4.1 8B	$1.30/mo
Nemotron 3 Nano 30B	$1.60/mo
Amazon Nova Lite	$1.92/mo

How AIStackPicker checks it

The cheapest build is recomputed from live data, not hand-picked. AIStackPicker ranks every verified, compatible model by full-stack cost on the cheapest always-on host that meets the runner's memory floor, and shows the source and age behind each price.

Open this build in the Builder Not sure? Answer 4 questions