Pricing

Verified output adds no separate fee. You pay only for the model usage it routes through the maxmodel.com gateway — metered on your existing token. The verification step itself, the deterministic substring check, is free.

How billing works

• A verified call routes ≥ 1 model call through the gateway (the EXTRACT step). That model usage is billed exactly like any other gateway call, on the same token — usage.modelCalls shows how many.
• The VERIFY step adds no charge — it’s local deterministic code: no second model, no extra round-trip.
• retry and verified.eval cost more only because they make more model calls (each retry, and each eval case, is another generation). See Limits & latency.
• For current per-model rates, see the gateway pricing at maxmodel.com.

Free to start

There’s nothing new to buy to use verified output:

• Use your existing maxmodel.com gateway token — there’s no separate “verified” key or plan.
• Add sources to a call you’re already making.
• Try it live in the Playground — no setup.

Enterprise

Building something high-stakes (legal, finance, healthcare)? Contact us to talk about:

• volume and committed-use pricing,
• a security / procurement review (see Data handling & privacy),
• requirements like data residency or a zero-retention route.

What you’re not charged for

• The verification check itself.
• A separate plan, seat, or key — it’s the same gateway token and the same quota.

A verification product shouldn’t nickel-and-dime the verification. You pay for model compute; the guarantee is free.