# I ran 120B MoE models on my 32GB M1 Max using under 16 GB

- HackerByte page: <https://hackerbyte.net/en/stories/a-user-shares-running-120b-parameter-moe-models-on-a-32gb-m1-max-with-low-memory-50430323>
- Original source: <https://mitln.github.io/moe-inference-engine/>
- Hacker News discussion: <https://news.ycombinator.com/item?id=50430323>
- Submitted to Hacker News: 2026-11-09T19:20:34.000Z

A user shares running 120B parameter MoE models on a 32GB M1 Max with low memory usage \(\<16GB\).

## AI-generated summary

No summary is available in this language yet.