Gemini API Billing Glitch Charges Users 27,000 Yuan in 12 Hours; Google Faces Refund Backlog

According to Beating, Gemini API developers have reported unauthorized charges triggered by two critical billing bugs in Google's AI platform. One developer was charged approximately 27,000 yuan within 12 hours, with some incidents involving zero output from the model. The issues stem from a "phantom cache" bug where deleted context caches continue generating charges at thousands of yuan per hour, and a "reasoning loop" bug that causes infinite token consumption reaching 64,000 tokens despite returning no usable output. Google's billing system experiences 32-to-72 hour processing delays and lacks automatic spending caps, preventing users from detecting fraudulent charges before funds are deducted. Multiple developers have since disabled Gemini's caching and reasoning features in production environments.
Disclaimer: The information on this page may come from third-party sources and is for reference only. It does not represent the views or opinions of Gate and does not constitute any financial, investment, or legal advice. Virtual asset trading involves high risk. Please do not rely solely on the information on this page when making decisions. For details, see the Disclaimer.
Comment
0/400
No comments