Corion

Pricing and credits

Commercial terms in this page are an MVP preview, not a binding quote. Per-token prices and provider markups must be published only after upstream model pricing is verified.

Prepaid usage

Corion uses prepaid credits. A successful model request deducts its metered cost from the account balance. A virtual key receives a spending budget based on the balance available when the key is created.

The planned checkout presets are $20, $50, and $100 credit top-ups. Their appearance in a deployment depends on Stripe configuration. Buying a pack adds usage credit; it does not guarantee a request rate, context window, model uptime, or expiration policy.

Token pricing

Prepaid credits are spent at the model-specific rates below. A request is billed for its metered input and output tokens, and there is no minimum charge per request.

ModelInput (per 1M tokens)Cached input (per 1M tokens)Output (per 1M tokens)
kimi-k3$2.50$0.25$12.50
deepseek-v4-pro-0813$0.435$0.0435$0.87
deepseek-v4-flash-0731$0.11$0.011$0.25

Input tokens served from the provider's prompt cache are billed at the cached rate, 10% of the standard input rate. Cached tokens are reported as usage.prompt_tokens_details.cached_tokens on every response and are excluded from the count billed at the standard input rate. Whether a request hits the cache is decided by the upstream provider; Corion does not guarantee a hit rate.

The dashboard's usage ledger is the source of truth for the MVP, subject to reconciliation with gateway spend records.

Before accepting live payments, Corion must publish:

  • the rounding and minimum-charge rules,
  • taxes, refunds, expiration, and failed-request treatment,
  • the currency used for balances and settlement.

Design partners

Teams evaluating sustained or specialized workloads can use the design partner form. Capacity, support, credits, and case-study terms are agreed individually; none are implied by submitting the form.

Enterprise

Dedicated capacity, private networking, data residency, contractual SLAs, and self-hosted deployment are roadmap options. They are not included in the Phase 1 service unless confirmed in a signed agreement.