Considered Gear

Free local AI starter guide

What AI can my computer run?

Find your first model.
Find out if you need to buy anything.

Tell us what you want to do and what you have. Get a model to try, a useful test prompt, and a next step that fits your situation.

For chat, document excerpts and coding help. No account. No price hunting.

Your starting plan

Find a starting point before you spend

The guided tool needs JavaScript. Read two starting plans below for a model, setup commands and a useful first test. For other hardware, use our model-memory guide.

Two starting plans you can use now

Prefer to read a plan first? Pick a matching example below. These are example profiles, not a scan of your computer or a measured fit. Both start with chat and writing on hardware you already own.

Install or update Ollama and check its current system requirements first.

16 GB Apple silicon Mac

First candidate: Qwen3.5 4B — exact tag qwen3.5:4b-q4_K_M.

Apple silicon Mac · 16 GB system unified memory. 3.3 GB is the published download size; runtime memory is higher.

After installing Ollama, open Terminal or PowerShell and run:

ollama run qwen3.5:4b-q4_K_M

Inside the chat, enter:

/set parameter num_ctx 4096

Try one useful task

Rewrite this as a friendly two-sentence message. Keep the date and time unchanged: Our appointment is on Tuesday at 3 pm. Please bring your laptop.

Check the answer: It should keep Tuesday and 3 pm, mention the laptop, and use two sentences. Then try one of your own everyday writing tasks.

Before buying: Use what you own. If the first try is slow or runs out of memory, reduce context or test a smaller model before spending.

RTX 3060 12 GB PC with 16 GB RAM

First candidate: Qwen3.5 9B — exact tag qwen3.5:9b-q4_K_M.

PC with an NVIDIA RTX GPU · 16 GB system RAM · 12 GB dedicated graphics memory. 6.6 GB is the published download size; runtime memory is higher.

After installing Ollama, open Terminal or PowerShell and run:

ollama run qwen3.5:9b-q4_K_M

Inside the chat, enter:

/set parameter num_ctx 4096

Try one useful task

Rewrite this as a friendly two-sentence message. Keep the date and time unchanged: Our appointment is on Tuesday at 3 pm. Please bring your laptop.

Check the answer: It should keep Tuesday and 3 pm, mention the laptop, and use two sentences. Then try one of your own everyday writing tasks.

Before buying: Use what you own. If the first try is slow or runs out of memory, reduce context or test a smaller model before spending.

These are starting experiments, not guaranteed results. Free memory, drivers and other applications affect performance. See the selection method and sources, checked 2026-10-18.

Useful advice, with the reasoning attached

This guide suggests a first experiment, not a guaranteed fit or a hardware performance ranking. The choices stay in your browser; the guide does not scan your computer, run a model or send your answers to an AI service.

How the starting model is chosen

For Apple silicon, our starting screen is 2B at 8 GB, 4B at 16 GB, and 9B at 24 GB or more unified memory. For NVIDIA RTX PCs with at least 16 GB system RAM, it is 4B at 6 GB VRAM for short chat (8 GB for document/code trials), and 9B at 12 GB VRAM or more. Other supported computers start with a 2B CPU trial; GPU acceleration is unconfirmed. Below 8 GB system RAM, or with unknown required memory, this guide asks for a different starting point or more information.

These are editorial capacity screens with room beyond the download, not measured free-memory requirements. We start chat at 4K context and document/code trials at 8K; model accuracy and speed must be checked with your own task. Images, full repositories and entire document libraries are outside this guide’s fit estimates. Spending preferences change buying guidance, not the hardware you actually have.

We use explicit Qwen3.5 model variants for a reproducible starting point. This is not a claim that they are the best models for every person. A larger budget never changes a “try what you own” answer into an automatic purchase.

Sources and last check

Checked 2026-10-18. Software support and model files can change. Check the linked requirements before installing or buying.

Ollama’s setup documentation covers the OS and driver requirements. Model pages supply the exact tags and published download sizes. Our memory thresholds are planning judgments; we do not claim measured speed on untested hardware.

Need more depth? Understand model memory · Read the hardware buying guide · Check ownership cost against a monthly AI bill.