How-to

How to draft shell scripts with the Nano tier

Last updated 2026-06-18 · Outlier v1.11.469

Quick answer

Lightweight zsh and bash one-liners generated locally on machines with as little as 6 GB of unified memory. The whole sequence below stays on the Mac.

What you need first for “how to draft shell scripts with the nano tier”

Apple Silicon Mac, macOS 12 or later, the unified-memory minimum that the chosen tier requires (6 GB for Nano, 12 GB for Lite, 24 GB for Core / Code / Vision, 32 GB for Plus). Internet is required only for the one-time model download.

Steps

  1. Switch to Nano. Settings > Models > Nano. The download is 2.37 GB.
  2. Ask in plain English. “A zsh function that resizes every PNG in a folder.”
  3. Read it before running it. Local generation is fast; verification is still on you.
  4. Pin the result. Outlier’s pinned-snippet panel keeps the working version visible.

What is the specific thing to know about this guide?

The Nano tier is 2.37 GB on disk and runs at 71.7 tok/s on the M1 Ultra reference, with an operator-reported ~32 tok/s on a 16 GB M4 Air (single-prompt observation, not yet σ-qualified).

What can go wrong with this guide?

How does this guide differ from the cloud equivalent?

Hosted shell-script tools are usually overkill for a 4B-class job. Nano on a 16 GB Mac handles this turn-by-turn at roughly 32 tok/s.

The Nano tier is 2.37 GB on disk and runs at 71.7 tok/s on the M1 Ultra reference, with an operator-reported ~32 tok/s on a 16 GB M4 Air (single-prompt observation, not yet σ-qualified).

What does this guide not claim about “how to draft shell scripts with the nano tier”?

The Nano tier is 2.37 GB on disk and runs at 71.7 tok/s on the M1 Ultra reference, with an operator-reported ~32 tok/s on a 16 GB M4 Air (single-prompt observation, not yet σ-qualified).

This guide does not claim feature parity with cloud-side workflows for “how to draft shell scripts with the nano tier”. Specifically, the product surface in v1.11.469 covers chat, file attachment, the local agent loop, project scoping, and the model picker. Cross-device sync, team workspaces, and shared session history are out of scope and are not on the v1.9 backlog either.

How fast is Nano on a 16 GB M4 Air?

Operator-reported observation: roughly 32 tok/s on a 16 GB M4 Air, 2026-05-04, single-prompt single-turn. This is not yet a σ-qualified number and does not appear in FINAL_LAUNCH_NUMBERS.md; treat it as indicative, not as a promise. The M1 Ultra reference is 71.7 tok/s on the same model.

Where does this guide fit in the rest of the lineup?

Nano’s 32K default context is more than enough for shell-script generation. The 6 GB unified-memory minimum means even an entry-level M1 Air handles this; the M4 Air is comfortable headroom.

One unique number

4 steps, zero network requests after the model is downloaded. Nano is not a code-tier; for shell scripts and small zsh utilities its fluency is enough, but for full programs the Lite or Core tier is the right choice.

Download Outlier for Mac

Requires Apple Silicon (M1, M2, M3, or M4) — Intel Macs are not supported. macOS 12+.

Outlier runs entirely on your Mac. No prompts leave the device. macOS 12+ on Apple Silicon (arm64). Apache 2.0 model weights. Back to home.