Coding on Mac mini M6, 32 GB
Apple M6. 32 GB. 170 GB/s.
Source: Apple Mac mini shop, 15 Oct 2026
Coding
Runs wellWell below DeepSeek (July 2026)
This compares a compressed local file (usually 4-bit, a smaller copy of the full model) with the full cloud model from July 2026. May be outdated.
A fast specialist. Only a small slice works on each word, so it feels quick. Newer Qwen3.6 builds exist; this is the one with a public Arena score.
Mac mini M4 Pro 48 GB (discontinued)
Around $2,199 · discontinued, used price · estimate
Source: Apple Newsroom, 22 September 2026, plus the last listed configuration, 15 Oct 2026
Install guides from the makers
- AnythingLLM's docs · AnythingLLM · Windows, Mac, Linux
- GPT4All's site · GPT4All · Windows, Mac, Linux
- Jan's documentation · Jan · Windows, Mac, Linux
- KoboldCpp's site · KoboldCpp · Windows, Mac, Linux
- llama.cpp's site · llama.cpp · Windows, Mac, Linux
- LM Studio's app guide · LM Studio · Windows, Mac, Linux
- LocalAI's site · LocalAI · Windows, Mac, Linux
- Msty's site · Msty · Windows, Mac, Linux
- Ollama's download page · Ollama · Windows, Mac, Linux
- Open WebUI's docs · Open WebUI · Windows, Mac, Linux
- Pinokio's site · Pinokio · Windows, Mac, Linux
- text-generation-webui's site · text-generation-webui · Windows, Mac, Linux
Not into local AI? Try these instead
Ordinary apps for code & tech. These are normal links to the maker's site.
#ad Shop links go straight to the store. No referral code is switched on yet, so those links do not pay LocalAIBuddy. A link is marked #ad only after that program is live. Affiliate disclosure. Prices checked 14 Oct 2026. Apple links on this page are not paid.
Fit rule checked 14 Oct 2026. canirun.ai.