I am using the Gemini API via OAuth 2.0 authentication with the following scopes: https://www.googleapis.com/auth/userinfo.email , https://www.googleapis.com/auth/userinfo.profile , openid , https://www.googleapis.com/auth/cloud-platform , and https://www.googleapis.com/auth/generative-language.retriever . Despite using OAuth access tokens (not API keys), I am still incurring charges when calling the generateContent endpoint (https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash:generateContent ) from a GCP project with Gemini API enabled. I want clarity on whether using OAuth scopes should still result in billing under my project or if there is an option to use a personal quota to avoid charges for small-scale testing. I also came across another scope, ./auth/generative-language.peruserquota , which claims to “Use Gemini models with your personal quota,” but it is marked as deprecated and lacks proper documentation. What is the official status of this scope, and what is the recommended approach for developers who want to test Gemini API without billing their GCP project? Clear documentation on the difference between OAuth usage and project-based billing for Gemini API is urgently needed because this is causing confusion and unexpected costs.
Hello,
Welcome to the Forum!!
For billing related issues I would recommend raising your concern at billing support.