The short version
A local chat app for running open models on modest Mac hardware.
From the HackerLinks archive
A local chat app for running open models on modest Mac hardware.
The short version
A local chat app for running open models on modest Mac hardware.
Why it caught our attention
A commenter reported GPT-4-level local use on a 16GB MacBook Air.
Where it surfaced on Hacker News
Editorial paraphrase
The discussion linked Samosa Chat alongside a claimed 7–9 tokens/sec Qwen setup.
Original threadRunning Gemma 4 26B at 5 tokens/sec on a 13-year-old Xeon with no GPU