Inference now runs in every region

Open models now run next to your functions in all 38 regions, billed by the token.

Calls stay inside the region, so prompts and answers never cross the internet on their way to a model.

Start with three model sizes today. More arrive every month.

All Articles