What's on your mind?
Every reply is generated by a real language model running on the devices of volunteers somewhere in the world, no datacenter involved.
No GPU in the pool can currently hold Llama 3.2 3B.
Chat resumes the moment a capable worker comes online — or add yours. Open the worker page in another tab, let the model load, and come back — your GPU will be serving this very page.
Become a worker →