Billing and credits

Availability

Billing is open.

Read the terms of use for payment and credit terms.

HTTP 402 insufficient_credits uses permission_error. Add funds before starting a new request. Read balance_cents, required_cents and add_funds_url in error.details.

Text prices follow price card H (November 11, 2026).

Cached input costs less.

Text caching is automatic. An optional prompt_cache_key improves reuse within a workspace. Cache lifetime, minimum size and eviction are not published. See Text caching for prices, usage fields and the console rate.

Refused questions are not charged. An all-refused request has no text, question, file or call charge.

An abstained answer under a request policy is charged like any answered question.

Charged once per answered call; never on refusals, zero-answer, free, promotional or replay calls.

[ Answered call ]
   |
   |-- text charges
   |
   |-- answered questions
   |
   |-- call fee
   |
   '-- applicable file fees

[ Entire call refused ]
   |
   v
[ No charge ]

See Inputs for inline formats, prices and availability.

Add text, answered-question and answered-call charges once per call.

PDF file fees cover ingestion; extracted text is charged at the text rate.

Text examples count total billed input, including the rendered questions and schema.

Each USD 1 paid buys a credit. The minimum purchase is USD 5. New buyers can add up to USD 500 a purchase and USD 1,000 a UTC day, per person and per workspace. Starting credit is subject to eligibility checks. Once per eligible canonical verified email address, on your first workspace, after eligibility and verification checks. Credit may be withheld or placed under review. Purchased credit expires 12 months after purchase. Your starting credit does not expire.

Credits API

[ Credit balance ]
        |
        v
[ Subtract active holds ]
        |
        v
[ Available credit ]

Read your credit balance with GET /v1/credits. See Credit balance for the reply fields and read limit.

Usage limits

Workspace usage limits still apply. Read Rate limits and use the response's retry and reset values to pace work.