Show XO: Open-source 'Strata' to run 1250 billion parameter AI model locally on gaming PCs (github.com)

Writing language: Korean Read in the original language

- Strata is a free open-source project enabling the large AI model with 1250 billion parameters (Qwen3.8-Flash-Next) to run on typical gaming PCs running Windows or Linux with NVIDIA or AMD graphics cards having at least 12GB VRAM. - It requires a minimum of 32GB RAM and about 80GB of storage. During installation, it automatically detects the graphics card and recommends appropriate model size and settings, supporting download and execution. - The model overcomes memory limits by distributing load across GPU, RAM, CPU, and SSD. It reads long texts at a rate of over 1,000 tokens per second and generates responses at 53 to 93 tokens per second depending on GPU performance (benchmarked by the development team on two gaming PCs). - It supports integration with other apps and coding agents via an OpenAI-compatible API and runs locally so that data is not sent to external clouds.

Found on

GitHub ↗ / 7,673 stars(checked 2026-10-03 22:59 KST)

1 comment

Editorial opinion startupxo 1d ago

There's a notice saying that during the first installation, the RAM from 35 to 55 GB may be locked for a long time and the PC could freeze. Since it requires a heavy memory load during the initial loading phase before it becomes usable, whether users can tolerate this will be a key factor in deciding to adopt local execution. It would be good to experience this initial process considering your PC's performance and your work patterns.

Writing language: Korean

Related reads

1 comment