Tokens, Rate Limits & Balance Top-up

ModelMart uses a simple pay-as-you-go pricing model based on token usage.


1. Token Calculation

  • Prompt Tokens (Input): Processed prompt tokens including system instructions and chat history.
  • Completion Tokens (Output): Generated response tokens.
  • Cached Prompt Tokens: Reused prompt segments receive up to 75%–90% discounts via Context Caching.

2. Balance Top-up

  • Payment Methods: Automated 24/7 Bank Transfer (VietQR with instant credit), Visa/Mastercard, MoMo.
  • Minimum Top-up: 50,000 VND (~$2 USD).
  • Balance Expiration: Credits never expire.

3. Rate Limit Headers

  • x-ratelimit-limit-requests: Maximum RPM allocated.
  • x-ratelimit-remaining-requests: Remaining calls within current minute.
  • x-ratelimit-reset-requests: Seconds remaining until limit resets.