← sources

M6 Mac Mini (32GB RAM) for Local LLM?

source: reddit (reddit:selfhosted)published 2026-09-25 00:22ingested 2026-09-26 02:36

retained text

I'm considering getting an M6 Mac Mini (32GB RAM) to run headless with Ollama, for local LLM for my self-hosted apps. I currently use OpenRouter for most of this, but really want to stop sending anything out of my network.

These are the main apps I'd like to use this for:

* Frigate (Generative AI Enrichments)
* Home Assistant (Whisper-STT/Piper-TTS/Home Assistant Voice/AI Tasks)
* Linkwarden (Tagging)
* Mealie (Ingredient parsing, recipe import)
* Papra (Tagging)
* Securo (Chat about accounts)

Is this a reasonable use case? Would something like gemma4:26b be a good fit for this, and would it run reasonably fast on this hardware?

Thanks in advance!

document fetched 2026-09-26 00:30 · canonical https://reddit.com/r/selfhosted/comments/1wpi2r6/m6_mac_mini_32gb_ram_for_local_llm