Pricing

Data Safety API subscriptions.

Monthly CPEN API credit, with per-request spend itemized in your dashboard. DS2-DS4 add 1% of the model usage charge after cache reconciliation; DS0 and DS1 add none.

Card billing checkout availableMonthly usage credit

Monthly subscription plans

Pick a monthly credit level.

Plans renew monthly; contact billing support to cancel before renewal. Processor billing appears when available; taxes may apply.

Monthly usage plan
Payment method
Have a referral code? "aimaster" grants 1.5x included usage on this checkout.
$9/moStarter$9.00 included CPEN API usage credit120 RPM · 5 API keys
  • Evaluation credit
  • Small workloads
Recommended
$19/moBuilder$19.00 included CPEN API usage credit600 RPM · 30 API keys
  • Monthly account credits
  • Billing + referral support
$49/moGrowth$49.00 included CPEN API usage credit1200 RPM · 50 API keys
  • Team account credits
  • Route health visibility
$99/moScale$99.00 included CPEN API usage credit1800 RPM · 75 API keys
  • Production account credits
  • Billing controls
$199/moMax$199.00 included CPEN API usage credit2400 RPM · 100 API keys
  • Highest public allowance
  • 100 API keys

Sign in, choose a plan, and apply any eligible referral code before billing starts.

Data Safety gateway guardrails

The model usage charge is calculated from measured input and output tokens, then prompt-cache savings are reconciled. DS2-DS4 add 1% of that cache-reconciled model usage charge; DS0 and DS1 add none.

Plan billingMonthly

Renews monthly; cancel through your payment processor or contact support.

Included usage1:1 USD

Plan dollars become same-value account credits.

Policy controls+1% DS2-DS4

DS2-DS4 add Data Safety enforcement to the request.

Model usage rates

Dashboard spend and API settlement use these public CPEN rates with a 1,000-token minimum billable unit for any nonzero request, then apply prompt cache reconciliation and the DS2-DS4 fee when applicable.

GPT 5.4 Minicpen/gpt-5-4-mini
Input$0.75 / 1M tokens
Output$4.50 / 1M tokens
GPT 5.6 Lunacpen/gpt-5-6-luna
Input$1.00 / 1M tokens
Output$6.00 / 1M tokens
GPT 5.6 Terracpen/gpt-5-6-terra
Input$2.50 / 1M tokens
Output$15.00 / 1M tokens
Billing minimum

Every nonzero request has a 1,000-token minimum.

CPEN keeps the prompt/completion ratio when scaling a short request up to the minimum billable unit. Requests above 1,000 tokens use their actual prompt and completion tokens.

Short example on cpen/gpt-5-4-mini: 42 prompt + 18 completion tokens are billed as 700 input + 300 output tokens, so the model charge is
$0.00187500
Regular example on cpen/gpt-5-4-mini: 1,699 prompt + 46 completion tokens are above the minimum, so the model charge is
$0.00148125

Payment processor

Card billing is available for signed-in accounts after processor and CPEN risk checks.

Payment processor
Card billing
Status
Checkout eligibility confirmed

Credits are account-bound, non-transferable, and usable only inside CPEN.

Billing is handled by the processor when enabled. CPEN handles product and billing support.

Billing support: support@cpenrouter.space

TermsRefund policyAcceptable usePrivacy

Prompt cache billing reconciliation

CPEN reserves uncached input cost first, then reconciles cached-token savings after the provider response.

Frequently asked questions

What am I paying for?
Monthly CPEN API credit. DS2-DS4 add 1% of the model usage charge after cache reconciliation; DS0 and DS1 add none.
Why is there a 1% Data Safety fee?
It covers classification, redaction, block/deny checks, and audit metadata. The fee basis is the model usage charge after cache reconciliation.
Why can a short request cost more than a direct per-token estimate?
Any nonzero request uses a 1,000-token minimum billable unit before cache reconciliation and any DS2-DS4 fee.
Can I get a refund?
See the refund policy for eligibility and how to request one.