I tried turning my old Samsung Galaxy into a local LLM-hosting machine, and it worked better than I expected
I had to use several workarounds, though.
A Samsung Galaxy S23 FE user experimented with running a large language model (LLM) on their old device, expecting it to perform poorly due to limited memory and processing power. They installed PocketPal, a client application, but encountered memory issues with larger models like Gemma 4 E2B, causing the app to crash. Switching to Termux and the llama-cpp package allowed them to successfully load and run Gemma 4 E2B, achieving usable performance with split-screen mode.
While the model generated inaccuracies, it performed well in tasks like summarization and email replies, proving to be a viable alternative to cloud-based AI services for daily use.
Written by urgent.news from XDA Developers's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
