Open-source Strata runs DeepSeek locally on an RTX 3060 at 9-11 tokens/s

Strata, a free MIT-licensed desktop app for Windows/Mac/Linux requiring no account, lets users run LLMs locally—including DeepSeek V4.1 on an RTX 3060 at 9-11 tokens per second—by paging model weights from SSD when they exceed available memory.

2026-11-08 ~ 2026-11-08 · 2 related posts