OpenAI RPM, TPM, and TPD Limit Management in Production Systems
Manage four independent rate limits or watch production fail at 2am.
Sofía Reyes-Montiel
Contributing Editor, AI Economics
Sofía Reyes-Montiel
Sofía holds a graduate degree in computational economics from Universidad Autónoma de Madrid and has been covering pricing models in emerging technology markets since 2014, first for a European trade publication and later independently. Her work at AI Spend Weekly focuses on the structural economics of token-based billing and how model providers design pricing to influence developer behavior.
5 stories
KV Cache and Prompt Caching Across OpenAI, Anthropic, and Google
Different providers require different strategies to actually save money on prompt caching.
Sofía Reyes-Montiel
Shared Credentials and the Spend Attribution Breakdown Problem
Shared API keys hide which teams actually spend money on AI models.
Sofía Reyes-Montiel
Multimodal Token Accounting for Vision and Audio Inputs
Understanding how images and audio convert to tokens reveals hidden cost differences of up to 9x.
Sofía Reyes-Montiel
Input vs Output Token Pricing Across Major Providers
Output tokens cost 2-8x more than input.
Sofía Reyes-Montiel