GoodPut AI

Billing and recharges

Usage-based billing, time-based prices, balance recharges, and bills.

Billing rules

Usage is billed by token, at the current catalog prices (¥ / million tokens). Each request is metered and priced across four dimensions:

DimensionDescription
InputPrompts, conversation history, tool definitions, and other request input
OutputGenerated content, including reasoning tokens
Cache readCached prompt tokens, priced well below ordinary input; see prompt caching
Cache writeSeparately priced for some models; otherwise charged at the input price
  • Streaming requests are billed on tokens actually generated.
  • Check each request’s actual charge in request logs.
  • Some models show a crossed-out price and a discounted price; the discounted price is charged.

Time-based prices (peak and off-peak)

Some models have higher peak-time prices and lower prices at other times.

  • See each model’s time-based pricing chart, or the public platform API, for windows and prices.
  • The request’s start time determines its pricing window, in Beijing time. A long request crossing a window boundary is not split.
  • Schedule batch and offline work in off-peak windows to lower costs.

Long-context pricing tiers

Some large-context models price each request by its total input volume:

  • Tier volume includes cache reads and cache writes.
  • On reaching a tier’s starting point, such as ≥ 200,000 tokens, the entire request uses that tier’s prices. Below all tier thresholds, standard prices apply.
  • See the Long-context Pricing Tiers table on the model page, or price.tiers[] in the public API.
  • Time windows and context tiers can combine: select the time window, then the tier. Top-level tiers[] contains standard-window tier prices; windows[].tiers contains each window’s tier prices. Request charge details identify the selected window and tier.

Balance and recharges

  • Finance → Add Funds supports Alipay and WeChat Pay, credited immediately. Enterprises can fund organization balances by bank transfer; see enterprise bank transfers.
  • Recharged funds do not expire.
  • View balances, orders, and the ledger under Finance. Programs can use an API Key for balance queries.
  • When funds run out, requests return 429. The OpenAI-compatible code is credit_balance_exhausted; Anthropic’s error.type is billing_error. See errors. Adding funds restores service automatically; keys need not be recreated.

Budgets and caps

The platform provides several levels of cost control for teams and enterprises:

LevelControls
OrganizationDefault member budget, automatically applied to new members
TeamTeam budget and separate request/token rate limits
API KeyPer-key cap; requests rejected when reached

See organizations and teams for configuration.

Plans and subscriptions

Finance → Plan Management offers recharge packages and subscriptions. Benefits, deduction rules, and overage handling are defined by each plan.

Bills and invoices

  • In an organization, Finance → Cost Invoices summarizes spending by billing period for reconciliation.
  • Contact us for a VAT invoice, providing the organization name and invoicing details.

Enterprise price agreements

Enterprise customers can sign a fixed-price agreement. Prices remain fixed during the contract term, without fixing usage volume. Contact sales for details.

On this page