No cloud needed: Nvidia creates gaming-centric AI chatbot that runs on your GPU Nvidia G-Assist has limited functionality right now, but it runs locally.

Google’s Gemma 3 is an open, single-GPU AI with a 128K context window Gemma 3 is optimized to run on almost anything, from powerful GPUs to a smartphone.

DeepSeek's distilled new R1 AI model can run on a single GPU, putting it within reach of hobbyists.