Pay-as-you-go: Slower response times and unconsumed credit balance with gemini-3.6-flash

Google Cloud Build with AI Generative AI & Foundational Models
1

Hi everyone,

I’m using the Gemini API on a Pay-as-you-go plan (Tier 1) via Google AI Studio with gemini-3.6-flash. Over the past few days, I’ve run into two issues:

  1. Increased Latency: Responses from gemini-3.6-flash are noticeably slower than usual. The Time to First Token (TTFT) and overall generation speed have significantly degraded compared to previous performance under the same workload.

  2. Credit Balance Not Decreasing: My credit balance in the AI Studio “Billing” tab hasn’t moved despite ongoing requests. I double-checked that my API keys belong to the correct GCP project with an active billing account linked.

Is anyone else experiencing increased latency on gemini-3.6-flash recently? Also, should the credit balance in the AI Studio tab reflect real-time API consumption, or is usage only tracked in the GCP Cloud Billing reports?

Thanks!