Skip to content
NEW GPT-5.5 / Claude 4.6 is now live — try it today →
Get Started Getting Started 5 min read

Billing & Balance

  • Pay per token: each model has a flat input and output price per 1M tokens.
  • Pay per call: image and video models are billed per request.
  • No plans, no monthly fees: pure usage-based billing with no minimums.
  • Failed requests are not billed.
  • Set a budget cap in the console to avoid surprise bills; the API returns 402 when your balance is exhausted and resumes automatically after you top up.
  • Check real-time usage and balance in the console.