Get started
Compute

Choose compute

Pick a workload and what matters most. Only AI inference is available today; the other workloads are on the way.

Workload

Priority

Price and speed together

FiltersOptional

Source

Run by

Your choice

AvailableBalanced

AI inference

Open models behind one API that OpenAI, Claude and Ollama clients already speak: chat, embeddings, rerank, images and audio.

Billed per token at the rates on the pricing page. Pricing

Create an API keyRead the quickstartOpen the playground

Serving today

Silicon Network operators

No Mac is online yet. Operators are joining, and requests can be served once the first ones are online. Network map

Model to send

silicon/auto:balanced

Today this is the same choice as Cheapest: the cheapest capable model, from a quality floor up.

How the model is chosen

Planned sources

  • Decentralized networks
  • GPU clouds and marketplaces
  • Public clouds