Run a 125-billion-parameter AI model on your own gaming PC. NVIDIA or AMD graphics card (12 GB or more), Windows or Linux, free and open source.
Strata runs Qwen3.8-Flash-Next on a normal PC. This is a large, smart AI model that usually needs a server. It chats, writes code, reads pictures and works with your apps and coding agents. Nothing leaves your PC.
We measured it on two ordinary gaming PCs. RTX 5070 (12 GB) writes answers at 94 tokens/s with Q2_0, reads your prompt at 2,650 tokens/s. RX 9070 XT (16 GB) writes at 60 tokens/s, reads at 1,160 tokens/s. An RTX 3090 (24 GB) should write about 100-140 tokens per second.
Strata is free. It needs a graphics card with 12 GB VRAM or more, 32 GB RAM (64 GB recommended), 80 GB disk, and Windows 10/11 or Linux. Install by downloading Strata and double-clicking START-HERE.bat on Windows or running ./setup.sh on Linux. Your browser opens the Strata app at http://127.0.0.1:8080.
Source: Hacker News · Summarized by HeadlinesBriefing