PRICING FAQ

    Commercial questions, answered precisely.

    Answers reflect confirmed website and product behavior. Anything not published is marked for console or sales confirmation.
    How are API tokens accounted for?

    Hosted API usage is measured using model input and output tokens. Rates vary by published model.

    Is cached input priced separately?

    Cached-input rates are not published in the connected InferX pricing source. Confirm support and pricing in the model platform before use.

    How does Monthly Access work?

    Monthly Access is $10 per month and includes a fixed amount of usage credits each billing cycle. The current included amount is published on the main Pricing page. After included credits are used, you can continue with pay-per-usage billing.

    How do invoices and taxes work?

    Invoice timing, tax treatment, and billing documentation are not published on the current website. Confirm these details through the billing portal or with InferX.

    What is the refund policy?

    No general refund policy is published in the connected pricing source. Billing adjustments require InferX confirmation.

    Are there minimum commitments?

    No commitment is published for pay-per-token endpoints. Sovereign and platform commitments depend on the deployment and require confirmation.

    How are Sovereign Endpoints™ billed?

    Billing depends on the model, GPU type, GPU count, capacity, and endpoint configuration.

    How does platform licensing work?

    Platform licensing is scoped around capacity, topology, deployment requirements, support, and professional services.

    What support is included?

    Support levels vary by product and commercial agreement. Platform and enterprise support require a scoped discussion.