Pricing

Transparent, usage-based pricing — these are the live rates every request is billed at.

Pay per token

You pay per million input/output tokens, debited from a prepaid balance — no subscriptions, no minimums. Top up any time in Billing.

Routing you control

Requests route to the least-loaded healthy node that serves your model — the rate is per model, the same wherever it lands. Pin a key to a region — EU, US, or Asia — when data residency matters.

Owners earn

Every request is split with the node owner who served it, paid in cash — owners keep 70% of the billed amount under the default rate plan. Run a node and the same balance flows back to you.

Model rates

Per 1M tokens in USD. The same numbers the meter bills with; see Models for context windows and capabilities.

Loading…

How billing works

Prepaid, metered, no surprises.

1 · Create an API key in Build → API keys.

2 · Add prepaid credit in Build → Billing.

3 · Call /v1/chat/completions — usage is metered per token and shown in Activity.