Run DeepSeek Locally on Your MacBook Pro

DeepSeek is open-weights reasoning and coding you can run on your own Mac — from a 4 GB R1 distill on a MacBook Air to the V4-Flash flagship on a Studio, with chain-of-thought that streams separately on the standard OpenAI wire. Here's the size that fits your Mac, and the one command that serves it.

DeepSeek is the open-weights family that put strong reasoning within reach of local hardware. The R1 line thinks step-by-step before it answers; the Coder models are tuned for code; and the V4-Flash flagship is one of the most capable open models you can run at all. On a Mac, rapid-mlx serves every one of them behind a drop-in OpenAI API — reasoning and all.

This guide is the short path: pick the right DeepSeek for your Mac's RAM, serve it with one command, and point your tools at it.

The 60-second version

curl -fsSL https://rapidmlx.com/install.sh | bash   # or: brew install rapid-mlx
rapid-mlx serve deepseek-r1-8b-4bit                 # pick your size below

Then talk to it — it's a standard OpenAI-compatible server on http://localhost:8000:

curl http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model": "default", "messages": [{"role": "user", "content": "A train leaves at 3pm going 60mph. Another leaves at 4pm going 80mph. When does the second catch the first?"}]}'

First run downloads the weights; after that, startup is seconds. No account, no API key, nothing leaves your machine.

Reasoning, coding, or the flagship?

DeepSeek covers three distinct jobs:

• R1 (reasoning) thinks before it answers — ideal for math, logic and

multi-step problems. A distilled 8B fits a small Mac; a 32B brings much stronger reasoning to a 24 GB+ machine.

• Coder is tuned for programming. deepseek-coder-v2-lite-16b is a

mixture-of-experts model that activates only a fraction of its parameters per token, so it decodes fast for its size.

• V4-Flash is the flagship — one of the strongest open models available,

for Macs with a lot of unified memory.

Which size fits your Mac

Sizes below are what the weights occupy on disk (and roughly in unified memory once loaded). Leave headroom for your context window and the rest of macOS.

Not sure where you land? The hardware tier guide maps every Mac from an 8 GB Air to a 512 GB Studio to a recommendation, and models.rapidmlx.com lets you filter the whole catalog by your RAM.

Tool calling and thinking work out of the box

serve auto-configures the right chat template, tool-call parser and reasoning parser for every DeepSeek alias, and grammar-constrained tool calling is on by default — so function calls come back as valid JSON, not as a paragraph that almost parses. That's the boring-but-load-bearing part: it means DeepSeek-on-your-Mac slots into tools that expect OpenAI semantics without any glue code.

Wire it into your editor or agent

The server speaks the OpenAI protocol, so anything with a "custom base URL" box can use it:

• Claude Code — run the CLI

agent fully local against your DeepSeek.

• Cursor — chat, Ctrl-K and

composer against http://localhost:8000/v1.

• Codex CLI and

Aider — both speak the same OpenAI-compatible endpoint.

For reasoning-heavy work, the R1 models are the draw; for code, reach for the Coder or a general-purpose family like Qwen 3.6. All of them serve through the same endpoint, so switching is one command.

How fast is it?

We publish real numbers rather than vibes: our five-model MacBook Pro benchmark puts local models through coding, reasoning, tool-calling and memory tests, and the M3 Ultra benchmark covers the desktop end. Reasoning models spend tokens thinking, so expect them to talk to themselves before answering — the payoff is markedly better results on hard problems.

Frequently asked questions

Can I run DeepSeek locally on a Mac with 8 GB of RAM?

Yes — deepseek-r1-8b-4bit is 4.0 GB on disk and runs on an 8 GB Mac, bringing R1's step-by-step reasoning to a base Air. For stronger reasoning, 24 GB fits the 32B distill (deepseek-r1-32b-4bit); the V4-Flash flagship needs a Studio-class machine.

What's the difference between DeepSeek R1 and DeepSeek Coder?

R1 is a reasoning model — it thinks step-by-step before answering, which helps on math, logic and multi-step problems. Coder is tuned for programming. deepseek-coder-v2-lite-16b is a mixture-of-experts model that decodes fast for its size. Pick R1 for reasoning, Coder for code — or switch between them with one command.

Does DeepSeek work with Claude Code and Cursor locally?

Yes. rapid-mlx serves DeepSeek behind an OpenAI-compatible endpoint, and R1's chain-of-thought streams as reasoning_content on the standard wire, so any tool with a custom base URL box — Claude Code, Cursor, Codex, Aider — can point at http://localhost:8000/v1 and use it with no bridge and no plugin.

Is running DeepSeek locally actually free and private?

The weights are open, rapid-mlx is Apache 2.0, and inference happens entirely on your Mac — no account, no per-token billing, and your prompts and code never leave the machine. The only costs are disk space for the weights and the electricity your Mac was already using.